Proteomic Analysis of Nrf2 Deficient Transgenic Mice Reveals Cellular Defence and Lipid Metabolism As Primary Nrf2-Dependent Pathways in the Liver

JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631

available at www.sciencedirect.com

www.elsevier.com/locate/jprot

Proteomic analysis of Nrf2 deficient transgenic mice reveals cellular defence and lipid metabolism as primary Nrf2-dependent pathways in the liver

Neil R. Kitteringhama,⁎,1, Azman Abdullaha,b,1, Joanne Walsha, Laura Randlea, Rosalind E. Jenkinsa, Rowena Sisona, Christopher E.P. Goldringa, Helen Powellc, Christopher Sandersona, Samantha Williamsa, Larry Higginsd, Masayuki Yamamotoe, John Hayesd, B. Kevin Parka a MRC Centre for Drug Safety Science, School of Biomedical Sciences, University of Liverpool, Sherrington Buildings, Liverpool, Merseyside, L69 3GE, United Kingdom b Department of Pharmacology, Faculty of Medicine, National University of Malaysia, Jalan Raja Muda Abdul Aziz, 50300 Kuala Lumpur, Malaysia c AstraZeneca R&D Alderley Park, Safety Assessment UK, Mereside, Alderley Park, Macclesfield, Cheshire, SK10 4TG England d Biomedical Research Centre, Ninewells Hospital and Medical School, University of Dundee, Dundee DD1 9SY, Scotland, United Kingdom e Division of Medical Biochemistry, Tohoku University Graduate School of Medicine, Tohoku, Japan

ARTICLE INFO ABSTRACT

Article history: The transcription factor Nrf2 regulates expression of multiple cellular defence proteins − − Received 22 January 2010 through the antioxidant response element (ARE). Nrf2-deficient mice (Nrf2 / ) are highly Accepted 31 March 2010 susceptible to xenobiotic-mediated toxicity, but the precise molecular basis of enhanced toxicity is unknown. Oligonucleotide array studies suggest that a wide range of gene products Keywords: is altered constitutively, however no equivalent proteomics analyses have been conducted. To Nrf2 define the range of Nrf2-regulated proteins at the constitutive level, protein expression − − Transgenic profiling of livers from Nrf2 / and wild type mice was conducted using both stable isotope Liver labelling (iTRAQ) and gel electrophoresis methods. To establish a robust reproducible list of Protein expression Nrf2-dependent proteins, three independent groups of mice were analysed. Correlative iTRAQ network analysis (MetaCore) identified two predominant groups of Nrf2-regulated proteins. As Lipid metabolism expected, one group comprised proteins involved in phase II drug metabolism, which were down-regulated in the absence of Nrf2. Surprisingly, the most profound changes were observed amongst proteins involved in the synthesis and metabolism of fatty acids and other lipids. Importantly, we show here for the first time, that the enzyme ATP-citrate lyase, responsible for acetyl-CoA production, is negatively regulated by Nrf2. This latter finding suggests that Nrf2 is a major regulator of cellular lipid disposition in the liver. © 2010 Elsevier B.V. All rights reserved.

1. Introduction macromolecules and can lead to the development of diseases, such as cancer, neurodegenerative disorders and cardiovascular Exposure to electrophiles and reactive oxygen species (ROS) may disease [1–3]. To counteract the damage caused by electrophiles result in intracellular damage to proteins, DNA and other and ROS, higher animals have developed elaborate defence

⁎ Corresponding author. Department of Pharmacology & Therapeutics, University of Liverpool, PO Box 147, Merseyside, L69 3GE, UK. Fax: +44 151 794 5540. E-mail address: [email protected] (N.R. Kitteringham). 1 Joint first authors who contributed equally to this work.

1874-3919/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jprot.2010.03.018 JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631 1613 mechanisms [4,5], which include the coordinated induction of a DE gels were obtained from GE Healthcare UK (Little Chalfont, battery of genes encoding phase II detoxifying enzymes and Bucks, UK). 8-plex isobaric tags for relative and absolute oxidative stress inducible proteins [6,7].Itisnowwellestab- quantification (iTRAQ) protein labelling kit/reagents were lished that a principal regulator of the cellular defence response purchased from AB Sciex (Framingham, MA, USA). Sequencing is the transcription factor termed nuclear factor erythroid-2 related grade trypsin was obtained from Promega UK (Southampton, factor 2 (Nrf2) [8–15]. Nrf2 been shown to play a key role in the Hants, UK). All other reagents were obtained from Sigma transcriptional activation of multiple genes involved in cellular (Poole, Dorset, UK). defence against ROS and electrophiles, such as NAD(P)H: quinone oxidoreductase (NQO1) [16], glutathione S-transferases 2.2. Animal studies (GSTs) [17,18], glutamate-cysteine ligase [19],haemoxygenase-1 (HO-1) [20], thioredoxin [12],andferritin[21]. All experiments were undertaken in accordance with criteria Nrf2 deficient transgenic mice have provided the most outlined in a license granted under the Animals (Scientific informative integrated model in which to examine the role of Procedures) Act 1986, and approved by the Animal Ethics Nrf2 in regulating the defence response to chemical insults, Committees of the University of Liverpool. Generation of the particularly in the liver. Inactivation of the nrf2 gene results in no Nrf2 knockout mouse and genotyping of progeny have been obvious phenotypic changes (except in aging female animals, described elsewhere [26,28]. Male mice of approximately where autoimmune diseases have been observed [22]), indicat- 10 weeks of age were used throughout the study. Mice were ing that Nrf2 is not essential for normal growth and develop- housed at a temperature range of 19 °C–23 °C under 12-h light/ ment [23]. Several studies focussing on individual proteins have dark cycles and given free access to food and water. Animals shown that the presence of Nrf2 is essential for the enhanced were killed by exposure to a rising concentration of CO2 followed expression of several antioxidant response proteins following by cervical dislocation. Livers were removed and snap-frozen − administration of certain chemical inducers, however, constitu- immediately in liquid N2, before being stored at 80 °C. tive expression of the same genes is often unaffected or only Three groups of mice were used: for the first iTRAQ − − marginally reduced by deletion of the Nrf2 gene [9,11,14,23–29]. analysis (iTRAQ analysis 1), 8 mice (4 Nrf(+/+) and 4 Nrf( / )) Acute exposure of Nrf2-deficient mice to a range of toxic were used, for the second iTRAQ analysis (iTRAQ analysis 2), 12 − − chemical insults has been shown to result in enhanced toxicity mice (6 Nrf(+/+) and 6 Nrf( / )) and for the 2DE gel analysis − − compared to their wild type counterparts. Multiple studies show 8 mice were used (4 Nrf(+/+) and 4 Nrf( / )). a reduced resistance to hepatotoxicity induced by a wide range of compounds, including paracetamol [25,30], carbon tetrachlo- 2.3. iTRAQ labelling of liver homogenates ride [31],pyrazole[32],ethanol[33] and pentachlorophenol [34]. Two possible explanations exist for the reduction in Whole liver homogenates (75 μg protein) from Nrf2(+/+) and − − chemically-induced hepatotoxicity seen in each of these Nrf2( / ) (n=4), were prepared in TEAB/SDS. iTRAQ reagent studies: first, the lack of Nrf2 may abrogate the animal'sability labelling was then carried out according to the Applied to up-regulate defence proteins in response to the chemical Biosystems protocol for an 8plex procedure. Briefly, samples stimulus or, second, the enhanced toxicity may simply reflect a were denatured, reduced and capped with methylmetha- constitutive reduction in defence proteins due to the absence of nethiosulfate (MMTS), before overnight digestion with trypsin Nrf2. Clearly, these two possible mechanisms are not mutually and then labelled with isobaric tags. For the first three iTRAQ exclusive and each may contribute to a different degree, runs, Nrf2(+/+) samples were labelled with tags 113 to 116 while − − depending on the nature of the chemical insult. Nevertheless, Nrf2( / ) samples received the 117 to 121 tags. In the fourth it is important to understand the relative contribution of each experiment, the sample labelling was reversed such that the mechanism for a given hepatotoxin in order to translate wild type animals had the heavier tags and the null mice the information gained in animal studies into improved clinical lighter tags, in order to control for labelling bias. iTRAQ- management of drug- or chemical-induced toxicity in man. labelled peptides were then pooled and diluted to 4 mL with Oligonucleotide microarray analysis of Nrf2 null mice suggests 10 mM potassium dihydrogen phosphate/25% acetonitrile that Nrf2 may regulate more than 200 genes, either constitutively (ACN; w/v). The pH of the samples was adjusted to <3 using or following exposure to a known inducer [29,35,36]. However, it is phosphoric acid prior to fractionation on a Polysulfoethyl A now well recognized that transcriptional up-regulation does not strong cation-exchange column (200×4.6 mm, 5 μm, 300 Å; always equate to increased protein expression [37]. Until now, no Poly LC, Columbia, MD). A flow rate of 1 mL/min was applied equivalent proteomic analysis of Nrf2 null mice has been and peptides eluted by increasing the concentration of KCl in undertaken to substantiate the mRNA expression changes at the mobile phase to 0.5 M over 60 min. Fractions of 2 mL were the protein level. Here we report a global analysis of constitutive collected and were dried by centrifugation under vacuum hepatic protein expression in Nrf2 null and wild type mice [10]. (SpeedVac, Eppendorf).

2.4. Mass spectrometric analysis of iTRAQ samples 2. Materials and methods For LC-MS/MS analysis of iTRAQ labelled samples, each cation 2.1. Materials exchange fraction was resuspended in 120 μL 5% ACN/0.05% trifluoroacetic acid (TFA) and 60 μL were loaded on column. Protein assay kits were from Bio-Rad (Hemel Hempstead, Samples were analysed on a QSTAR® Pulsar i hybrid mass Herts, UK). Immobiline Dry Strips and associated buffers for 2- spectrometer (AB Sciex) and were delivered into the 1614 JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631 instrument by automated in-line liquid chromatography 3-[(3-Cholamidopropyl)-dimethylammonio]-1-propane sulfo- (integrated LCPackings System, 5 mm C18 nano-precolumn nate (CHAPS), 10 mM 1,4-dithiothreitol (DTT), 1 mM EDTA]. and 75 μm×15 cm C18 PepMap column; Dionex, California, The homogenate was sonicated for 30 s and centrifuged at USA) via a nano-electrospray source head and 10 μm inner 150000 g for 45 min. The supernatant was assayed for protein diameter PicoTip (New Objective, Massachusetts, USA). The content [38] and stored at −80 °C. Aliquots of samples containing precolumn was washed for 30 min at 30 μL/min with 5% ACN/ equal quantities of protein (0.5 mg) were diluted to 350 μLwith 0.05% TFA prior to initiation of the solvent gradient in order to rehydration buffer (9 M urea, 2% w/v CHAPS, bromophenol blue reduce the level of salt in the sample. A gradient from 5% ACN/ (trace), 2% v/v immobilized pH gradient (IPG) buffer, 0.28% w/v 0.05% TFA (v/v) to 60% ACN/0.05% TFA (v/v) in 70 min was DTT) and incubated overnight with nonlinear Immobiline applied at a flow rate of 300 nL/min. The MS was operated in DryStrips (18 cm; pH 3–10 non-linear) in a re-swelling chamber. positive ion mode with survey scans of 1 s, and with an MS/MS The samples were separated in the 1st dimension by isoelectric accumulation time of 1 s for the three most intense ions. focussing (IEF) for 25 h at a constant temperature of 20 °C to Collision energies were calculated on the fly based on the m/z achieve a total of 75 000 Vh (MultiPhor II, GE Healthcare UK, Little of the target ion and the formula, collision energy=(slope×m/ Chalfont, Bucks, UK). The IPG strips were then incubated with z)+intercept. The intercepts were increased by 3–5 V com- equilibration buffer (50 mM Tris, 6 M urea, 30% v/v glycerol, 2% pared to standard data acquisition in order to improve the w/v SDS, bromophenol blue (trace) containing 1% w/v DTT) for reporter ion intensities/quantitative reproducibility. 15 min followed by incubation in the same buffer with the DTT replaced by 2.5% w/v iodoacetamide for a further 15 min. The 2.5. iTRAQ data analysis strips were applied to the surface of 12% w/v SDS-PAGE gels and sealed with agarose [39]. The samples were subjected to Data analysis was performed using ProteinPilot software electrophoresis at 20 W/gel and 25 °C for approximately 3 h (Version 3, AB Sciex, Warrington, UK). The data were analysed (Ettan Dalt 12, GE Healthcare UK, Little Chalfont, Bucks, UK). The with a fixed modification of MMTS-labelled cysteine, biological gels were then stained with colloidal Coomassie blue. modifications allowed and with the confidence set to 10% to enable the False Discovery Rate to be calculated from screening 2.8. Image analysis of 2-DE gels the reversed SwissProt database. Ratios for each iTRAQ label were obtained, using a wild type mouse (WT mouse 1) sample Colloidal Coomassie blue stained 2D gels were scanned using a as the denominator. The detected protein threshold (“unused GS710 calibrated imaging densitometer (BioRad, Hemel Hemp- protscore (conf)”) in the software was set to 1.3 to achieve 95% stead, UK). TIFF images were generated and were analysed using confidence. ImageMasterTM 2D Elite software, version 4.01 (Amersham Pharmacia Biotech, Buckinghamshire, England). Altogether − − 2.6. Network analysis eight gels were analyzed (4 Nrf2( / ) and 4 Nrf2(+/+)). An objective strategy for quantitative comparisons between wild type and The accession numbers of the 108 proteins identified as null liver samples was adopted to exclude the possibility of bias, significantly different following Benjamini–Hochberg adjust- as follows. The gels were initially analysed using an automated ment for multiple comparisons (p≤0.2) were converted to procedure to identify spots. The authenticity and outline of each Entrez gene IDs using the Database for Annotation, Visualiza- spot was validated by eye and edited where necessary. In each tion and Integrated Discovery (DAVID) (http://david.abcc. case approximately 500 validated spots were recorded from ncifcrf.gov/conversion.jsp ) and analysed for evidence of each gel. Spot matching was accomplished initially by auto- network wide changes in cellular phenotype using MetaCore mated fitting of the spots, followed by manual seeding of from GeneGo Inc., an integrated manually curated knowledge remaining spots that failed to match by automated fitting. A database for pathway analysis of gene lists (http://www. background value was subtracted for each gel and the spot genego.com/metacore.php). volumes normalised against the total volume of all matched The gene list was analysed using the Pathway Maps tool, spots. For each spot, the ratio between its intensity and the sum which maps the genes listed to defined signalling pathways of all spot intensities in the gel (normalized spot intensity) was that have been experimentally validated and are widely determined and used for quantitative comparison. Visual and accepted. The proteins deemed Nrf2-regulated according to quantitative comparisons were only sought in spots that were the criteria defined above were compared against a background matched in all four gels for a given treatment group. file containing all of the identified proteins which had similarly been converted to a list of Entrez gene IDs using DAVID. The 2.9. Identification of proteins from 2DE gels p values generated by the software were used to determine the statistical significance of the pathways identified. The p value Protein spots of interest were excised from Colloidal Coomas- represents the probability that a particular pathway will be sie blue-stained 2DE gels by automated spot excision (Ettan represented by chance given the number of genes in the Dalt Spot Picker, Amersham Biosciences) and were subjected experiment and the number of genes in the pathway. to tryptic digestion [40]. Gel pieces were washed with 100 μlof

50% (v/v) ACN/50 mM ammonium bicarbonate (NH4CO3) 2.7. 2-DE of liver homogenates (pH 7.8) for 15 min and were dried by centrifugation under vacuum (SpeedVac, Eppendorf). The dried gel pieces were Mouse livers were weighed and 0.3 g of tissue was homogenized rehydrated with 4–10 μl of digestion buffer (5 ng/μl of modified in 1 mL lysis buffer [40 mM tris, 7 M urea, 2 M thiourea, 4% (w/v) sequencing grade trypsin in 50 mM NH4CO3)andwere JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631 1615 incubated overnight at 37 °C. The resulting peptides were derived by Nioi et al. [44]. In order to identify ARE-like sequences, extracted by the addition of 30 μl of 60% ACN/1% TFA and a matrix-based pattern matching method was performed using incubation for 5 min in a sonicating water bath at 20 °C. The the programme ‘patser’. A position-specific scoring matrix samples were briefly centrifuged and the supernatants were (PSSM) was created based on the core ARE (cARE) position- collected. A further 30 μl of 60% ACN/1% TFA was added to the specific probability matrix published by Nerland [45].Inorderto gel plug and the sample was sonicated for 5 min. The calculate the background base frequencies, A/T and C/G supernatants were pooled and dried by centrifugation under frequencies were determined within the upstream sequences vacuum. The peptides were resuspended in 10 μl of 5% ACN/ of all the genes interrogated and a mean value for each base was 0.05% TFA. 0.5 μl of the peptide mixture was spotted onto a 96- defined (A/T 0.26, C/G 0.24). The derived PSSM was then used to position stainless steel target and was mixed 1:1 with matrix scan each of the genes shown to be significantly different − − [10 mg/ml α-cyano-4-hydroxycinnamic acid in 50% ACN/ 0.1% between Nrf2(+/+) and Nrf2( / ) mice. TFA]. Peptide mass finger prints were obtained on a Voyager DE Pro MALDI mass spectrometer (AB Sciex) and the resulting 2.12. Statistical analysis mass lists searched against the NCBInr database using Mascot software (Matrix Science). Scores of greater than 75 were 2.12.1. iTRAQ data regarded as sufficient for identification. Each significant iTRAQ data for proteins within the 1% false discovery rate and identification was checked for consistency between its for which full quantification data were obtained, were isoelectric point, molecular mass and mobility on the 2DE statistically analysed within the R computational environment gel and, where possible, coincidence with a published mouse [46]. R is an open source software environment for statistical liver proteome 2DE database [41,42]. computing and graphics (http://www.r-project.org/). Normality of data and equivalence of variance across the data sets was 2.10. Western immunoblotting for ATP-citrate lyase assessed by Shapiro–Wilk and F-tests, respectively, and also by inspection of histogram plots for all proteins identified. Data Whole liver homogenate (25 µg of protein) was separated by were then analysed by t-test using the module multtest,a denaturing electrophoresis on a 10% polyacrylamide gel package designed for re-sampling based multiple hypothesis (ProtoGel acrylamide solution and buffers, using Tris-Gly- testing. Benjamini–Hochberg corrections for multiple compar- cine-SDS running buffer) and transferred to a nitrocellulose isons were performed on all raw p values generated [47]. membrane (GE healthcare). After transfer, a Ponceau Red stain Protein expression differences between wild type and Nrf2-null was used to ensure equal loading and then the membrane was mice giving a p value of <0.05 by t-test and a Benjamini– blocked using 10% milk in 1x TBS/0.1%Tween for 30 min at Hochberg value ≤0.2 were accepted for further correlative room temperature, before incubation with a rabbit monoclo- network analysis. The Benjamini–Hochberg cut-off was set at nal antibody to ATP citrate lyase (ab40793, Abcam plc, Cam- 0.2 to avoid the exclusion of correlated Nrf2-regulated proteins bridge, UK) at 1:2000 with 2% milk in 1xTBS/0.1%Tween at 4 °C through application of too stringent a correction for multiple overnight. The membrane was washed several times with testing in accordance with multivariate modelling approaches TBS-Tween and then incubated with the secondary antibody to account for potential confounders [47]. (peroxidise-conjugated goat anti-rabbit immunoglobulin G, 1:10000 in TBS-Tween containing 2% milk) for 1h at room 2.12.2. 2DE gel data temperature. Enhanced Chemiluminescence Plus (GE Health- Data are expressed as mean±SEM for at least four separate care) was used to visualise the level of protein-antibody experiments. All values were analysed for non-normality using complex. Band volume was measured by densitometry using the Shapiro–Wilk test. Normally distributed values were Biorad Quantity One 1D Analysis Software (BioRad). compared using Student's unpaired t-test whilst non-normal values were analysed using the Mann–Whitney test. These 2.11. Identification of antioxidant response elements in the statistical analyses were performed using the SPSS statistical promoter regions of Nrf2-regulated genes software package, version 12 (Chicago, IL, USA). Statistical significance was accepted at p values of <0.05. ARE consensus sequences were sought in the 5′-flanking regions upstream of all genes identified in the initial iTRAQ analysis (iTRAQ analysis 1) as being Nrf2-regulated (p<0.05, Student's t-test). Public domain software (Regulatory Sequence 3. Results Analysis Tools, http://rsat.ulb.ac.be/rsat/) provided by the Service − − de Conformation des Macromolécules Biologiques et de Bioinformatique 3.1. iTRAQ analysis of Nrf2(+/+) and Nrf2( / ) mouse at the University Libre de Bruxelles [43] was used. 5′-flanking liver proteins sequences (2000 bp upstream of the start codon) were retrieved directly from the ENSEMBL database from within the RSAT Two independent sets of mice were analysed using iTRAQ stable package. Promoter sequences were then interrogated for ARE or isotope labelling. For the first analysis (iTRAQ analysis 1), − − ARE-like sequences using both string-based and matrix-based samples from four Nrf2(+/+) and four Nrf2( / ) mice were analysed protocols. String-based analysis was carried out using the using 8-plex iTRAQ reagents and the entire analysis was programme ‘dna search’ available within the RSAT web resource. repeated on four separate occasions. For the second group of − − The search term used was RTGABNNNGCA (where R=G/C, B=G/ mice (iTRAQ analysis 2), six Nrf2(+/+) and six Nrf2( / ) mice were C/T and N=any nucleotide) based on the consensus sequence compared on a single occasion using three sets of 4-plex iTRAQ 1616 JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631 reagents. The second set of mice was thus used as a validation 3.1.2. iTRAQ analysis 2 cohort to challenge the reproducibility of the protein changes For iTRAQ analysis 2, each run represents a single analysis of a observed in the initial “training” set. In each case iTRAQ data different set of wild-type and Nrf2-null liver samples. In this were processed using Protein Pilot version 3, including FDR. case, 1070 proteins were initially identified with a FDR below 1%, Table 1 shows the numbers of proteins identified and quantified of which 628 were associated with full quantitative datasets within the two independent iTRAQ analyses. (Table 1). There was little variation between the three runs with respect to protein numbers identified, however the final number 3.1.1. iTRAQ analysis 1 of unique proteins quantified was slightly lower than in iTRAQ For iTRAQ analysis 1, each of the four runs represents a full analysis 1. Following Student's t-test analysis with adjustment repeat analysis of the same mouse liver sample. Thus, each for multiple testing by Benjamini–Hochberg analysis, thirty protein expression value derived from iTRAQ analysis 1 eight proteins were found to be differentially regulated between − − represents the mean from four animals repeated on four Nrf2(+/+) and Nrf2( / ) liver samples, as shown in Table 4. occasions. In total, 1109 unique proteins were identified in at least one of the four runs within the FDR of 1% (Table 1); of these, 3.1.3. Cellular defence and lipid metabolism are the primary 769 proteins had complete data sets in at least one run across all biochemical functions regulated by Nrf2 eight mice, and were consequently accepted for full quantitative The proteins identified as Nrf2-regulated by the two iTRAQ analysis. Considerable variation was seen between the four analyses were independently subjected to correlative network runs, with the first run in particular giving relatively low analysis. The 108 proteins obtained from the initial iTRAQ proteome coverage. Nevertheless, all four runs were included analysis were submitted to the PANTHER database for for statistical analysis in order to maximise the number of alignment to specific cellular pathways. Fig. 2 is a pie chart proteins to include in the network analysis. Following statistical indicating the pathways identified along with the percentage analysis, 108 proteins were found to be differentially expressed of the proteins corresponding to each fraction. The most − − between Nrf2(+/+) and Nrf2( / ) mouse livers using the criteria prominent class of proteins were those involved in lipid, fatty defined above and these are listed in Table 2. Fig. 1 shows a acid and steroid metabolism (18%). Other functional groupings volcano plot of the entire data set highlighting proteins whose included metabolism (protein, carbohydrate and amino acid), expression was significantly different (t-test p<0.05) between electron transport, immunity and defence, and transport. In wild type and Nrf2-null mice (open circles). Proteins that were order to mine further into the specific pathways influenced by significantly different by at least 20% are shown as filled circles. Nrf2, the 108 proteins identified as Nrf2-regulated from iTRAQ Approximately equivalent numbers of proteins were found to be analysis 1 were subjected to pathway analysis using the − − up-regulated in Nrf2( / ) mice as were down-regulated. Whilst software MetaCore. Of the 108 proteins, 104 were recognised those that were significantly less abundant in Nrf2 null animals by MetaCore and 68 had been mapped to pathways. The corresponded mainly to proteins involved in phase II and phase significant proteins were analysed against a background file III drug disposition, in line with previous oligonucleotide array containing all proteins quantified across the four replicate and immunoblotting experiments, those that were up-regulat- runs. Of the 769 proteins in the background file, 752 were ed were mostly involved with lipid metabolism. Proteins whose recognised by the software and 504 had been mapped to function is identified within the Uniprot database (http://www. pathways. Ten pathways were identified as significantly uniprot.org/) as lipid metabolism or lipid transport are summa- different (p<0.05) between wild type and Nrf2 null mice, as rized in Table 3. shown in Table 5: seven of these are involved in fatty acid metabolism or other lipid-related processes. In particular, several of the pathways are associated with peroxisomes, Table 1 – Total numbers of proteins identified and suggesting that non-mitochondrial fatty acid metabolism may quantified with a false discovery rate (FDR) exclusion of 1% in iTRAQ analyses 1 and 2. be a specific target for Nrf2-associated protein expression. Amongst the ten significant pathways identified only one, the iTRAQ LC-MS No. of No. of proteins No. of glutathione metabolism pathway, is directly involved in analysis analysis proteins identified proteins identified above 1% quantified cellular defence against reactive oxygen species or electro- global FDR philes. Nevertheless, this pathway was populated by five differentially regulated proteins (Table 5). 1 Run 1 486 265 162 Run 2 1287 911 620 Run 3 1003 759 593 3.1.4. ATP-citrate lyase is negatively regulated by Nrf2 in mice Run 4 726 563 426 Since a role for Nrf2 as a negative regulator of proteins Total 1654 1109 769 involved in lipid metabolism has only recently been suggested [48,49], an attempt was made to verify some of the changes 2 Run 1 1068 825 654 observed by immunoblotting. One of the most significant Run 2 1065 780 661 differences observed in the experiments in iTRAQ analysis 1 Run 3 1068 711 637 involved ATP-citrate lyase. This showed a mean 1.75-fold Total 1717 1070 628 increase but in the “test” cohort a value of 1.2-fold was seen, Numbers are given for proteins identified with a confidence greater which failed to reach significance (data not shown). Conse- than 90% and for those characterized by at least 2 peptides. The quently, a comparison between the wild type and knockout number of proteins quantified relates to those proteins determined animals was conducted by Western immunoblotting in order to in all eight mouse liver samples. validate the original iTRAQ observation. Due to the difficulty in Table 2 – Nrf2-regulated mouse hepatic proteins identified in iTRAQ analysis 1. Relative expression compared to WT 1

− − Nrf2(+/+) Nrf2( / ) Fold change − − SwissProt Name n Average Average Mouse Mouse Mouse Mouse Geometric Lower Upper Mouse Mouse Mouse Mouse Geometric Lower Upper Nrf2( / ) BH acc. no. no. of coverage WT1 WT2 WT3 WT4 mean 95% 95% KO1 KO2 KO3 KO4 mean 95% 95% Nrf2(+/+) p peptides (%) CI CI CI CI

P02762 Major urinary protein 6 4 19.8 54.9 1.00 1.35 1.29 1.54 1.28 1.07 1.53 0.47 0.28 0.48 0.64 0.45 0.32 0.63 0.35 0.057 P17427 AP-2 complex subunit alpha2 1 1.0 2.5 1.00 1.25 1.93 1.51 1.38 1.05 1.82 0.46 0.43 0.71 0.66 0.55 0.43 0.71 0.40 0.064 P10649 Glutathione 4 13.8 39.2 1.00 1.31 1.00 1.11 1.10 0.97 1.24 0.47 0.53 0.44 0.42 0.46 0.42 0.51 0.42 0.009 S-transferase Mu 1 Q61656 Probable 1 2.0 5.4 1.00 1.31 1.19 1.35 1.20 1.05 1.38 0.37 0.85 0.51 0.79 0.59 0.40 0.87 0.49 0.148 ATP-dependent RNAhelicaseDDX5

Q91WG8 Bifunctional UDP-N- 1 2.0 4.0 1.00 0.98 1.13 1.19 1.07 0.98 1.17 0.49 0.60 0.68 0.59 0.59 0.52 0.67 0.55 0.022 1612 (2010) 73 PROTEOMICS OF JOURNAL acetylglucosamine 2-epimerase P19157 Glutathione S-transferase P 1 4 43.0 76.3 1.00 1.21 0.94 1.12 1.06 0.95 1.19 0.62 0.56 0.60 0.54 0.58 0.55 0.62 0.55 0.011 P17717 UDP- 4 5.8 15.5 1.00 1.16 0.99 1.08 1.05 0.98 1.13 0.59 0.57 0.56 0.61 0.58 0.56 0.61 0.55 0.004 glucuronosyltransferase 2B5 Q63836 Selenium-binding protein 2 4 26.0 47.9 1.00 1.26 0.99 1.48 1.17 0.96 1.41 0.61 0.59 0.67 0.72 0.65 0.59 0.71 0.55 0.051 Q8VCC2 Liver carboxylesterase 1 3 2.3 4.6 1.00 1.34 1.06 0.94 1.08 0.93 1.25 0.62 0.60 0.58 0.70 0.62 0.58 0.68 0.58 0.042 Q60991 Cytochrome P450 7B1 1 2.0 7.1 1.00 1.43 1.66 1.65 1.40 1.11 1.77 0.85 0.82 0.82 0.79 0.82 0.80 0.84 0.58 0.073 P46425 Glutathione 1 39.0 71.0 1.00 0.70 0.76 0.61 0.75 0.61 0.93 0.47 0.45 0.43 0.43 0.44 0.43 0.46 0.59 0.063 S-transferase P2 P24472 Glutathione 2 2.5 17.6 1.00 1.01 0.99 0.92 0.98 0.94 1.02 0.49 0.62 0.76 0.50 0.58 0.48 0.72 0.60 0.073 S-transferase A4 O35660 Glutathione 1 7.0 24.3 1.00 0.68 0.67 0.89 0.80 0.66 0.97 0.50 0.69 0.42 0.40 0.49 0.38 0.62 0.61 0.179 S-transferase M6 P00186 Cytochrome P450 1A2 3 3.0 10.9 1.00 1.14 1.26 1.21 1.15 1.04 1.27 0.59 0.61 0.91 0.86 0.73 0.58 0.91 0.63 0.186 Q9EQU5 Protein SET 1 1.0 6.2 1.00 1.22 1.34 0.99 1.13 0.97 1.31 1.05 0.63 0.57 0.71 0.72 0.56 0.94 0.64 0.199 Q91X77 Cytochrome P450 2C50 3 6.0 16.5 1.00 1.30 1.29 1.33 1.22 1.07 1.40 0.67 0.67 1.03 0.87 0.80 0.65 0.98 0.65 0.162 Q6XVG2 Cytochrome P450 2C54 4 3.5 8.5 1.00 1.00 0.96 1.04 1.00 0.97 1.03 0.54 0.70 0.77 0.77 0.69 0.58 0.81 0.69 0.090

Q91XE8 Transmembrane 2 1.5 11.4 1.00 0.67 0.70 0.60 0.73 0.58 0.91 0.49 0.47 0.49 0.57 0.50 0.46 0.55 0.69 0.153 – 1631 protein 205 P15105 Glutamine synthetase 4 9.8 25.1 1.00 1.16 1.06 1.29 1.12 1.01 1.25 0.70 0.67 0.99 0.83 0.79 0.66 0.93 0.70 0.182 O55060 Thiopurine 2 1.0 5.4 1.00 0.85 0.99 0.75 0.89 0.78 1.02 0.49 0.71 0.71 0.70 0.65 0.54 0.77 0.72 0.194 S-methyltransferase O35490 Betaine-homocysteine 4 18.3 45.2 1.00 0.81 1.11 1.12 1.00 0.87 1.16 0.76 0.67 0.78 0.80 0.75 0.70 0.81 0.75 0.148 S-methyltransferase 1 P24549 Retinal dehydrogenase 1 4 13.8 31.2 1.00 1.07 1.10 1.22 1.10 1.01 1.19 0.80 0.76 0.84 0.92 0.83 0.77 0.90 0.76 0.127 P06801 NADP-dependent 3 8.0 20.5 1.00 1.32 1.16 1.22 1.17 1.04 1.31 0.75 0.93 1.06 0.84 0.89 0.77 1.02 0.76 0.201 malic enzyme P62858 40 S ribosomal protein S28 4 1.0 17.4 1.00 1.03 1.08 1.11 1.05 1.01 1.10 0.87 0.76 0.81 0.82 0.82 0.77 0.86 0.77 0.038 Q91VA0 Acyl-coenzyme A 3 6.3 20.4 1.00 0.95 1.03 0.90 0.97 0.91 1.03 0.80 0.71 0.75 0.75 0.75 0.72 0.79 0.78 0.039 synthetase ACSM1, mitochondrial Q9JIF7 Coatomer subunit beta 2 3.0 3.9 1.00 0.86 0.92 0.94 0.93 0.87 0.99 0.77 0.74 0.65 0.74 0.72 0.67 0.77 0.78 0.044 O55125 Protein NipSnap homolog 1 3 1.0 4.0 1.00 0.76 0.80 0.93 0.87 0.77 0.98 0.62 0.68 0.68 0.73 0.68 0.64 0.72 0.78 0.201 Q99JI4 26 S proteasome 2 1.0 3.9 1.00 0.76 0.75 0.74 0.81 0.70 0.93 0.65 0.67 0.60 0.61 0.63 0.60 0.67 0.78 0.182

non-ATPase regulatory 1617 subunit 6 (continued(continued on next page) 1618

Table 2 (continued) Relative expression compared to WT 1

Q99J99 3-mercaptopyruvate 2 2.0 10.9 1.00 0.99 0.87 0.94 0.95 0.89 1.01 0.72 0.71 0.80 0.79 0.75 0.71 0.80 0.79 0.057 sulfurtransferase Q76MZ3 Serine/threonine- 2 1.0 3.4 1.00 0.76 0.94 0.83 0.88 0.78 0.99 0.65 0.73 0.60 0.82 0.70 0.61 0.80 0.80 0.204 protein phosphatase 2A 65 kDa regulatory subunit A alpha Q9Z0X1 Apoptosis-inducing 2 1.0 2.1 1.00 0.99 0.89 0.86 0.93 0.87 1.00 0.67 0.78 0.76 0.80 0.75 0.69 0.81 0.80 0.114 factor 1, mitochondrial 1612 (2010) 73 PROTEOMICS OF JOURNAL O70475 UDP-glucose 3 5.7 19.5 1.00 0.94 0.96 1.06 0.99 0.94 1.04 0.77 0.87 0.81 0.75 0.80 0.75 0.85 0.81 0.061 6-dehydrogenase Q8R1G2 Carboxymethylene- 2 2.5 12.9 1.00 0.85 0.96 1.06 0.96 0.88 1.06 0.71 0.75 0.82 0.86 0.78 0.72 0.85 0.81 0.178 butenolidase homolog Q8VCU1 Liver carboxylesterase 31-like 3 10.0 20.4 1.00 0.95 0.90 0.83 0.92 0.85 0.99 0.74 0.69 0.79 0.78 0.75 0.70 0.79 0.81 0.117 Q8VCA8 Secernin-2 1 1.0 4.0 1.00 1.09 1.05 0.92 1.01 0.94 1.09 0.87 0.76 0.93 0.78 0.83 0.76 0.91 0.82 0.156 Q91VS7 Microsomal glutathione 4 5.0 30.2 1.00 0.92 0.95 0.95 0.95 0.92 0.99 0.84 0.71 0.73 0.85 0.78 0.71 0.86 0.82 0.162 S-transferase 1 Q9D6Y7 Peptide methionine 3 3.0 16.7 1.00 1.08 1.04 1.06 1.04 1.01 1.08 0.83 0.82 0.93 0.85 0.86 0.81 0.91 0.82 0.044 sulfoxide reductase P70441 Na(+)/H(+) exchange 3 1.7 5.7 1.00 0.82 0.84 0.73 0.84 0.74 0.96 0.75 0.68 0.69 0.68 0.70 0.67 0.73 0.83 0.198 regulatory cofactor NHE-RF1 Q8VCW8 Acyl-CoA synthetase 3 6.3 18.5 1.00 0.98 1.00 0.93 0.98 0.94 1.01 0.81 0.77 0.85 0.82 0.81 0.78 0.85 0.83 0.030 family member 2, mitochondrial P57776 Elongation factor 1-delta 3 3.7 23.5 1.00 0.90 0.87 0.83 0.90 0.83 0.97 0.84 0.76 0.78 0.74 0.78 0.74 0.82 0.87 0.180 P07759 Serine protease inhibitor 3 9.0 24.5 1.00 1.03 1.16 1.05 1.06 1.00 1.13 0.91 0.91 0.87 0.97 0.92 0.88 0.96 0.87 0.123 A3K –

Q91ZJ5 UTP-glucose-1-phosphate 3 3.0 8.0 1.00 0.99 1.08 1.06 1.03 0.99 1.08 0.86 0.89 0.96 0.88 0.90 0.85 0.94 0.87 0.156 1631 uridylyltransferase P11352 Glutathione peroxidase 1 4 4.5 26.1 1.00 0.96 1.07 1.12 1.04 0.97 1.11 0.90 0.88 0.95 0.94 0.92 0.89 0.95 0.89 0.193 P60867 40 S ribosomal protein S20 3 2.3 16.2 1.00 0.91 1.01 0.91 0.96 0.90 1.01 0.90 0.81 0.86 0.84 0.85 0.82 0.89 0.89 0.178 Q9JII6 Alcohol dehydrogenase 4 5.5 25.0 1.00 0.95 0.97 0.89 0.95 0.91 1.00 0.85 0.89 0.86 0.84 0.86 0.84 0.88 0.90 0.121 [NADP+] Q9DBJ1 Phosphoglycerate mutase 1 2 8.0 44.3 1.00 0.98 1.02 1.04 1.01 0.98 1.03 1.09 1.06 1.04 1.10 1.07 1.05 1.10 1.07 0.128 Q8BVI4 Dihydropteridine reductase 3 3.0 18.4 1.00 1.09 1.05 1.09 1.06 1.02 1.10 1.14 1.12 1.15 1.16 1.14 1.12 1.16 1.08 0.178 Q8BH00 Aldehyde dehydrogenase 3 13.3 31.9 1.00 1.07 1.08 1.14 1.07 1.02 1.13 1.19 1.19 1.14 1.15 1.17 1.14 1.20 1.09 0.206 family 8 member A1 Q8BFR5 Elongation factor Tu, 3 2.7 10.4 1.00 0.93 0.97 0.91 0.95 0.91 0.99 1.05 0.99 1.07 1.05 1.04 1.00 1.08 1.09 0.144 mitochondrial Q3UQ44 Ras GTPase-activating- 3 3.3 3.4 1.00 1.04 0.99 0.96 1.00 0.97 1.03 1.11 1.15 1.09 1.08 1.10 1.08 1.13 1.11 0.057 like protein IQGAP2 P21107 Tropomyosin alpha-3 chain 1 1.0 3.5 1.00 1.00 1.03 1.05 1.02 1.00 1.04 1.12 1.23 1.04 1.16 1.14 1.06 1.22 1.12 0.188 Q64374 Regucalcin 4 13.8 42.6 1.00 1.07 1.02 1.08 1.04 1.01 1.08 1.12 1.22 1.24 1.09 1.17 1.10 1.24 1.12 0.148 P45952 Medium-chain specific 3 4.7 14.2 1.00 1.06 1.11 1.08 1.06 1.02 1.11 1.20 1.20 1.24 1.13 1.19 1.15 1.23 1.12 0.095 acyl-CoA dehydrogenase, mitochondrial P62991 Ubiquitin 4 4.8 50.3 1.00 1.08 1.03 1.15 1.06 1.00 1.13 1.17 1.27 1.18 1.15 1.19 1.14 1.24 1.12 0.178 Q8CHT0 Delta-1-pyrroline-5- 3 7.0 18.0 1.00 1.05 1.09 0.97 1.03 0.98 1.08 1.17 1.20 1.23 1.05 1.16 1.09 1.24 1.13 0.193 carboxylate dehydrogenase, mitochondrial Q99J08 SEC14-like protein 2 3 6.3 25.5 1.00 1.08 1.12 1.21 1.10 1.02 1.19 1.27 1.18 1.32 1.23 1.25 1.19 1.31 1.14 0.186 Q02053 Ubiquitin-like modifier- 4 3.5 5.6 1.00 1.05 1.00 1.04 1.02 1.00 1.05 1.10 1.19 1.24 1.14 1.16 1.11 1.22 1.14 0.073 activating enzyme 1 O88569 Heterogeneous nuclear 4 4.5 12.5 1.00 0.98 1.03 1.08 1.02 0.98 1.07 1.22 1.17 1.10 1.16 1.16 1.12 1.21 1.14 0.090 ribonucleoproteins A2/B1 Q9QXD6 Fructose-1,6-bisphosphatase 4 16.0 46.5 1.00 0.93 0.98 0.92 0.96 0.92 0.99 1.16 1.14 1.05 1.01 1.09 1.02 1.16 1.14 0.144 Q99JI6 Ras-related protein Rap-1b 2 1.0 6.5 1.00 1.02 0.99 1.01 1.00 1.00 1.01 1.21 1.20 1.08 1.10 1.14 1.08 1.21 1.14 0.121 P50580 Proliferation- 2 2.0 6.6 1.00 1.01 1.07 0.98 1.01 0.97 1.05 1.10 1.19 1.25 1.10 1.16 1.09 1.23 1.14 0.127 associated protein 2G4 Q9R0Q7 Prostaglandin E synthase 3 2 1.5 13.1 1.00 1.05 1.11 1.10 1.06 1.01 1.11 1.17 1.14 1.21 1.36 1.21 1.12 1.31 1.14 0.199 Q9DCN2 NADH-cytochrome 3 6.0 27.9 1.00 1.00 0.98 0.98 0.99 0.98 1.00 1.18 1.05 1.11 1.21 1.13 1.07 1.21 1.15 0.072 b5 reductase 3

Q99LP6 GrpE protein homolog 1, 2 1.0 6.5 1.00 1.18 1.10 1.09 1.09 1.02 1.17 1.31 1.22 1.20 1.31 1.26 1.20 1.32 1.15 0.142 1612 (2010) 73 PROTEOMICS OF JOURNAL mitochondrial Q9JI75 Ribosyldihydronicotinamide 2 3.0 19.3 1.00 0.88 0.98 0.91 0.94 0.88 1.00 1.16 1.12 1.02 1.05 1.08 1.02 1.15 1.15 0.144 dehydrogenase [quinone] P00329 Alcohol dehydrogenase 1 4 13.0 32.8 1.00 1.04 1.10 1.05 1.05 1.01 1.09 1.26 1.20 1.18 1.20 1.21 1.18 1.24 1.16 0.039 P06151 L-lactate dehydrogenase 4 12.5 36.1 1.00 0.92 0.96 0.90 0.94 0.90 0.99 1.22 1.18 1.00 1.06 1.11 1.01 1.22 1.18 0.178 Achain Q8CHR6 Dihydropyrimidine 2 2.0 2.6 1.00 0.96 1.00 1.01 0.99 0.97 1.01 1.33 1.13 1.16 1.14 1.18 1.10 1.28 1.20 0.072 dehydrogenase [NADP+] P00405 Cytochrome c oxidase 2 2.5 15.2 1.00 1.16 0.98 0.99 1.03 0.95 1.11 1.19 1.26 1.18 1.28 1.23 1.18 1.28 1.20 0.105 subunit 2 Q9QXE0 2-hydroxyacyl-CoA lyase 1 3 2.7 7.5 1.00 1.02 0.89 0.83 0.93 0.84 1.02 1.10 1.17 1.13 1.07 1.12 1.08 1.16 1.20 0.117 Q60932 Voltage-dependent anion- 2 1.5 6.6 1.00 1.01 0.95 1.05 1.00 0.96 1.04 1.24 1.22 1.22 1.16 1.21 1.18 1.25 1.21 0.022 selective channel protein 1 Q61207 Sulfated glycoprotein 1 3 2.0 2.8 1.00 1.08 1.17 1.24 1.12 1.02 1.23 1.35 1.50 1.32 1.30 1.37 1.28 1.46 1.22 0.121 Q8VC12 Probable urocanate 4 8.0 14.9 1.00 1.20 1.12 1.09 1.10 1.02 1.18 1.36 1.39 1.30 1.33 1.35 1.31 1.39 1.23 0.063 hydratase P80316 T-complex protein 1 1 4.0 14.4 1.00 1.22 1.02 1.03 1.06 0.97 1.16 1.37 1.30 1.23 1.34 1.31 1.25 1.37 1.23 0.103 subunit epsilon

P50172 Corticosteroid 11-beta- 4 2.8 10.2 1.00 1.19 0.98 1.09 1.06 0.98 1.16 1.31 1.20 1.41 1.32 1.31 1.23 1.39 1.23 0.103 – 1631 dehydrogenase isozyme 1 Q8VCR7 Abhydrolase domain- 3 3.7 22.1 1.00 1.17 1.01 1.24 1.10 0.99 1.22 1.36 1.36 1.31 1.41 1.36 1.32 1.40 1.24 0.123 containing protein 14B Q9DD20 Methyltransferase-like 3 3.0 15.2 1.00 0.93 0.90 1.05 0.97 0.90 1.04 1.19 1.23 1.16 1.21 1.20 1.17 1.23 1.24 0.044 protein 7B Q61171 Peroxiredoxin-2 3 1.7 10.4 1.00 1.11 0.94 1.21 1.06 0.95 1.18 1.26 1.29 1.38 1.35 1.32 1.27 1.37 1.25 0.142 P24270 Catalase 4 12.3 25.9 1.00 1.25 1.02 1.14 1.10 0.99 1.21 1.39 1.33 1.41 1.40 1.38 1.35 1.42 1.26 0.096 P16460 Argininosuccinate synthase 4 26.8 47.6 1.00 0.82 1.02 0.89 0.93 0.84 1.03 1.26 1.31 1.03 1.11 1.17 1.05 1.31 1.26 0.162 P31786 Acyl-CoA-binding protein 4 4.8 39.1 1.00 0.85 0.92 0.83 0.90 0.83 0.97 1.22 1.20 1.05 1.09 1.14 1.06 1.22 1.26 0.073 Q61425 Hydroxyacyl-coenzyme A 3 3.3 12.1 1.00 1.04 0.98 1.04 1.01 0.99 1.04 1.09 1.25 1.51 1.31 1.28 1.12 1.46 1.26 0.156 dehydrogenase, mitochondrial A3KMP2 Tetratricopeptide repeat 3 2.3 6.4 1.00 0.95 0.96 1.16 1.02 0.93 1.11 1.21 1.41 1.24 1.33 1.29 1.21 1.39 1.27 0.117 protein 38 Q99PG0 Arylacetamide deacetylase 3 3.3 12.8 1.00 1.13 1.13 1.12 1.09 1.03 1.16 1.44 1.25 1.36 1.54 1.39 1.28 1.52 1.27 0.072 P12787 Cytochrome c oxidase 2 3.0 36.6 1.00 1.12 0.85 1.12 1.02 0.89 1.16 1.18 1.27 1.45 1.33 1.31 1.20 1.42 1.29 0.148

subunit 5A, mitochondrial 1619 (continued(continued on next page) 1620 Table 2 (continued) Relative expression compared to WT 1

P32020 Non-specific lipid-transfer 4 11.8 25.1 1.00 1.34 1.09 1.22 1.15 1.02 1.31 1.53 1.41 1.45 1.54 1.48 1.42 1.55 1.29 0.117 protein P55096 ATP-binding cassette 3 2.3 5.8 1.00 1.29 1.08 1.29 1.16 1.02 1.32 1.46 1.37 1.53 1.64 1.50 1.39 1.61 1.29 0.142 sub-family D member 3 P05201 Aspartate aminotransferase 3 5.7 19.4 1.00 0.83 1.02 0.90 0.94 0.85 1.03 1.34 1.39 1.04 1.12 1.22 1.06 1.39 1.30 0.178 cytoplasmic P19096 Fatty acid synthase 4 30.3 17.7 1.00 1.10 1.03 1.15 1.07 1.00 1.13 1.35 1.40 1.44 1.36 1.39 1.35 1.43 1.30 0.022

Q9R0H0 Peroxisomal 3 12.0 24.2 1.00 1.06 1.01 0.93 1.00 0.95 1.05 1.31 1.33 1.29 1.31 1.31 1.29 1.33 1.31 0.009 1612 (2010) 73 PROTEOMICS OF JOURNAL acyl-coenzyme A oxidase 1 P17665 Cytochrome c oxidase 1 2.0 47.6 1.00 0.89 0.87 1.07 0.95 0.87 1.05 1.34 1.15 1.34 1.26 1.27 1.18 1.36 1.33 0.072 subunit 7C, mitochondrial Q9QXF8 Glycine 4 18.0 47.6 1.00 1.27 1.23 1.42 1.22 1.06 1.41 1.63 1.66 1.60 1.61 1.63 1.60 1.66 1.34 0.117 N-methyltransferase P35492 Histidine ammonia-lyase 3 6.7 13.1 1.00 1.08 1.17 1.12 1.09 1.02 1.17 1.57 1.54 1.33 1.44 1.47 1.36 1.58 1.34 0.044 P83940 Transcription elongation 1 1.0 8.0 1.00 1.02 0.87 1.11 1.00 0.90 1.10 1.40 1.31 1.33 1.31 1.34 1.30 1.38 1.35 0.050 factor B polypeptide 1 P18242 Cathepsin D 2 5.0 17.7 1.00 1.20 1.00 1.42 1.14 0.97 1.35 1.49 1.86 1.47 1.48 1.57 1.40 1.75 1.37 0.178 P25688 Uricase 4 7.5 27.4 1.00 1.11 0.98 1.04 1.03 0.98 1.09 1.39 1.42 1.43 1.50 1.43 1.39 1.48 1.39 0.009 Q9QXD1 Peroxisomal acyl-coenzyme 1 2.0 3.8 1.00 1.38 1.28 1.32 1.23 1.07 1.42 1.68 1.73 1.57 2.01 1.74 1.57 1.93 1.41 0.117 Aoxidase2 P62984 60 S ribosomal protein L40 1 1.0 19.2 1.00 0.96 0.93 1.07 0.99 0.93 1.05 1.42 1.52 1.41 1.27 1.40 1.31 1.51 1.42 0.022 Q99P30 Peroxisomal coenzyme 4 4.5 30.3 1.00 1.14 0.90 1.14 1.04 0.93 1.16 1.60 1.45 1.40 1.48 1.48 1.40 1.57 1.43 0.044 A diphosphatase NUDT7 Q9DBM2 Peroxisomal bifunctional 4 2.3 4.7 1.00 1.35 1.10 1.20 1.16 1.02 1.31 1.52 1.76 1.60 1.73 1.65 1.54 1.76 1.43 0.057 enzyme O35423 Serine-pyruvate 3 1.0 3.1 1.00 0.88 0.86 0.93 0.92 0.86 0.98 1.55 1.67 1.15 1.25 1.39 1.17 1.65 1.51 0.066

aminotransferase, – mitochondrial 1631 Q8VBT2 L-serine dehydratase 3 4.3 22.3 1.00 0.72 0.97 0.90 0.89 0.77 1.03 1.47 1.58 1.18 1.20 1.35 1.17 1.55 1.51 0.078 Q8JZR0 Long-chain-fatty-acid-CoA 2 3.5 7.7 1.00 0.97 0.99 0.88 0.96 0.91 1.02 1.58 1.56 1.20 1.59 1.47 1.29 1.69 1.53 0.044 ligase 5 Q91V92 ATP-citrate synthase 3 11.3 14.4 1.00 1.13 1.05 1.09 1.07 1.01 1.12 1.97 2.02 1.84 1.79 1.90 1.80 2.01 1.78 0.003 P62827 GTP-binding nuclear 1 1.0 8.8 1.00 1.69 1.44 1.61 1.41 1.12 1.78 2.37 2.79 3.07 2.11 2.56 2.17 3.01 1.82 0.101 protein Ran P13516 Acyl-CoA desaturase 1 1 2.0 9.0 1.00 1.43 1.09 1.19 1.17 1.00 1.36 4.04 2.12 1.53 2.58 2.41 1.62 3.59 2.07 0.153 Q8VCH0 3-ketoacyl-CoA thiolase 3 6.7 25.2 1.00 1.89 1.43 1.44 1.41 1.09 1.82 2.92 3.61 4.04 2.98 3.35 2.88 3.91 2.39 0.044 B, peroxisomal Q05816 Fatty acid-binding 4 1.8 17.0 1.00 1.24 1.01 0.85 1.02 0.87 1.18 3.64 3.17 2.74 2.62 3.02 2.61 3.50 2.97 0.009 protein, epidermal

− − Relative expression of hepatic proteins in livers of Nrf2 wild type (Nrf2(+/+) ) and null (Nrf2( / )) mice determined in iTRAQ analysis 1. All values are expressed relative to a wild type control mouse (WT1). Proteins listed were significantly different in the null mice compared with wild type according to Student's t-test followed by Benjamini-Hochberg (BH) correction for multiple testing at a significance level of p≤0.2. Four replicate iTRAQ analyses were conducted on each sample and the number of runs in which each protein appeared is designated by n in column 3. The values for each mouse thus represent the average of n replicates. The fold change was calculated from the geometric mean values obtained from the 4 individual mice. Variance of the geometric mean for the four animals in each − − group is expressed as upper and lower 95% confidence intervals (CI). Proteins are listed according to their expression in Nrf2( / ) mice relative to wild type animals in ascending order of the fold-change value. JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631 1621

protein (Ponceau) stain was used as a loading control. The immunoblot (Fig. 3) confirmed that ATP-citrate lyase was indeed considerably over-expressed in Nrf2-null mouse liver with densitometric analysis indicating a 2.6-fold enhancement. As far as we are aware, this association between ATP-citrate lyase and Nrf2 has not been shown previously.

3.2. Identification of Nrf2-dependent liver proteins: 2-DE studies

Mouse liver protein extracts were separated by 2-DE and the protein spots were visualized following staining with colloidal Coomassieblue.Atotalof8gelswereproduced(4Nrf2(+/+) ,4 − − Nrf2( / )). Approximately 500 spots/gel were detected by the automated spot detection algorithm across the 8 gels. Using the criteria defined above (statistically significant difference within one or more comparisons and spot detected in all four gels for any Fig. 1 – Volcano plot of the entire set of proteins quantified treatment group) 8 spots were differentially expressed indicating during iTRAQ analysis 1. Each point represents the difference − − a role for Nrf2 in their regulation. These spots are labelled in the in expression (fold-change) between Nrf2(+/+) and Nrf2( / ) representative gel image (from a wild type control mouse liver) mice plotted against the level of statistical significance. Solid shown in Fig. 4a. Montage images of the differentially expressed lines represent differential expression differences of ±20% spots across the four treatment groups are shown in Fig. 4b. and a significance level of p<0.05 (Student's t-test). Proteins Table 6 lists all the gel spots whose intensity varied in one or more represented by diamonds were not differentially expressed. of the treatment groups. Proteins associated with each gel spot Circles represent proteins that gave a raw p value of <0.05 were identified by MALDI mass spectrometric analysis. One and Benjamini–Hochberg value of ≤0.2. protein (glutathione S-transferase pi) was identified in three of the differentially regulated spots. Attempts to relate the iTRAQ data with the 2DE gel protein identifying a suitable ‘housekeeping’ protein (we have found expression changes were hampered by the small number of both actin and GAPDH to be unreliable when comparing whole proteins identified by the gel-based approach. This may reflect − − liver homogenates from Nrf2( / ) and Nrf2(+/+) mice) the total the fact that Nrf2-regulated proteins have properties that are

Table 3 – Differentially up-regulated proteins listed in the UniProt database as involved in lipid synthesis or metabolism in ITRAQ analysis 1. SwissProt Name Subcellular Relative expression compared to Nrf2(+/+) mouse 1

acc. no. location − − Nrf2(+/+) Nrf2( / ) Fold change p − − Nrf2( / ) Geometric 95% CI Geometric 95% CI Nrf2(+/+) mean mean

Q05816 Fatty acid-binding protein, epidermal C 1.02 (0.87–1.18) 3.02 (2.61–3.50) 2.97 0.009 Q8VCH0 3-Ketoacyl-CoA thiolase B, P 1.41 (1.09–1.82) 3.35 (2.88–3.91) 2.39 0.044 peroxisomal P13516 Acyl-CoA desaturase 1 ER 1.17 (1.00–1.36) 2.41 (1.62–3.59) 2.07 0.153 Q91V92 ATP-citrate synthase C 1.07 (1.01–1.12) 1.90 (1.80–2.01) 1.78 0.003 Q8JZR0 Long-chain-fatty-acid-CoA ligase 5 ER, Mi 0.96 (0.91–1.02) 1.47 (1.29–1.69) 1.53 0.044 Q9DBM2 Peroxisomal bifunctional enzyme P 1.16 (1.02–1.31) 1.65 (1.54–1.76) 1.43 0.057 Q99P30 Peroxisomal coenzyme A P 1.04 (0.93–1.16) 1.48 (1.40–1.57) 1.43 0.044 diphosphatase NUDT7 Q9QXD1 Peroxisomal acyl-coenzyme A P 1.23 (1.07–1.42) 1.74 (1.57–1.93) 1.41 0.117 oxidase 2 Q9R0H0 Peroxisomal acyl-coenzyme A P 1.00 (0.95–1.05) 1.31 (1.29–1.33) 1.31 0.009 oxidase 1 P19096 Fatty acid synthase C 1.07 (1.00–1.13) 1.39 (1.35–1.43) 1.30 0.022 P32020 Non-specific lipid-transfer protein C 1.15 (1.02–1.31) 1.48 (1.42–1.55) 1.29 0.117 Q61425 Hydroxyacyl-coenzyme A Mi 1.01 (0.99–1.04) 1.28 (1.12–1.46) 1.26 0.156 dehydrogenase, mitochondrial P31786 Acyl-CoA-binding protein Mi 0.90 (0.83–0.97) 1.14 (1.06–1.22) 1.26 0.073 P50172 Corticosteroid 11-beta- ER 1.06 (0.98–1.16) 1.31 (1.23–1.39) 1.23 0.103 dehydrogenase isozyme 1 Q9QXE0 2-Hydroxyacyl-CoA lyase 1 P 0.93 (0.84–1.02) 1.12 (1.08–1.16) 1.20 0.117 1622 JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631

Table 4 – Nrf2-regulated mouse hepatic proteins determined in iTRAQ analysis 2 (test set). SwissProt Name Relative expression compared to Nrf2(+/+) mouse 1 Fold p

acc. no. − − change (+/+) ( / ) − − Nrf2 Nrf2 Nrf2( / ) (+/+) Geometric Lower Upper Geometric Lower Upper /Nrf2 mean 95% CI 95% CI mean 95% CI 95% CI

P10649 Glutathione S-transferase Mu 1 1.00 0.91 1.10 0.44 0.40 0.47 0.44 0.001 P17717 UDP-glucuronosyltransferase 2B5 0.99 0.93 1.06 0.55 0.52 0.57 0.55 0.001 Q8VCC2 Liver carboxylesterase 1 1.06 0.96 1.17 0.59 0.54 0.64 0.56 0.001 Q91X77 Cytochrome P450 2C50 0.97 0.87 1.08 0.56 0.46 0.69 0.58 0.001 P19157 Glutathione S-transferase P 1 0.95 0.89 1.01 0.58 0.54 0.63 0.62 0.001 Q9D379 Epoxide hydrolase 1 0.97 0.90 1.05 0.63 0.59 0.68 0.65 0.001 Q64458 Cytochrome P450 2C29 1.08 0.90 1.29 0.75 0.62 0.90 0.69 0.001 P30115 Glutathione S-transferase A3 1.03 0.98 1.08 0.72 0.66 0.78 0.70 0.001 P24549 Retinal dehydrogenase 1 0.94 0.86 1.04 0.68 0.58 0.80 0.72 0.021 O70475 UDP-glucose 6-dehydrogenase 1.09 0.97 1.23 0.79 0.63 0.99 0.73 0.183 Q62452 UDP-glucuronosyltransferase 1-9 0.99 0.93 1.06 0.73 0.58 0.92 0.74 0.183 Q91VA0 Acyl-coenzyme A synthetase ACSM1, 0.97 0.91 1.04 0.79 0.74 0.84 0.81 0.001 mitochondrial Q64442 Sorbitol dehydrogenase 1.02 0.92 1.13 0.84 0.78 0.91 0.83 0.081 P97494 Glutamate-cysteine ligase catalytic subunit 1.15 1.06 1.25 0.95 0.89 1.02 0.83 0.021 Q8CG76 Aflatoxin B1 aldehyde reductase member 2 1.06 1.01 1.11 0.88 0.81 0.96 0.83 0.013 Q9CQX2 Cytochrome b5 type B 1.01 0.91 1.13 0.85 0.76 0.94 0.83 0.197 Q9JII6 Alcohol dehydrogenase [NADP+] 1.01 0.98 1.04 0.86 0.80 0.92 0.85 0.003 Q8VCW8 Acyl-CoA synthetase family member 2, 1.00 0.93 1.08 0.86 0.79 0.93 0.86 0.132 mitochondrial O55022 Membrane-associated progesterone receptor 1.03 0.96 1.11 0.89 0.81 0.98 0.86 0.207 component 1 P47738 Aldehyde dehydrogenase, mitochondrial 1.02 0.99 1.06 0.88 0.85 0.92 0.86 0.000 Q8QZS1 3-hydroxyisobutyryl-CoA hydrolase, 1.07 1.00 1.13 0.94 0.89 1.00 0.89 0.084 mitochondrial Q9ET01 Glycogen phosphorylase, liver form 0.98 0.96 1.00 0.87 0.80 0.94 0.89 0.081 O35945 Aldehyde dehydrogenase, cytosolic 1 0.97 0.93 1.01 0.86 0.83 0.89 0.89 0.024 Q8VDJ3 Vigilin 1.10 1.05 1.15 0.99 0.94 1.04 0.90 0.069 Q9EQ20 Methylmalonate-semialdehyde 1.00 0.98 1.03 0.92 0.87 0.98 0.92 0.140 dehydrogenase [acylating], mitochondrial Q9Z2I8 Succinyl-CoA ligase [GDP-forming] subunit 1.02 0.99 1.06 0.96 0.92 0.99 0.93 0.121 beta, mitochondrial Q99P30 Peroxisomal coenzyme A diphosphatase 0.98 0.93 1.03 1.10 1.05 1.17 1.13 0.039 NUDT7 Q9CW42 MOSC domain-containing protein 1, 0.95 0.90 1.00 1.09 1.04 1.15 1.15 0.095 mitochondrial Q9QXD6 Fructose-1,6-bisphosphatase 1 1.02 0.96 1.09 1.21 1.10 1.33 1.18 0.117 P31786 Acyl-CoA-binding protein 1.01 0.94 1.07 1.19 1.06 1.34 1.18 0.207 P24369 Peptidyl-prolyl cis-trans isomerase B 0.95 0.89 1.01 1.12 1.05 1.20 1.18 0.017 Q8VDM4 26 S proteasome non-ATPase regulatory 0.90 0.80 1.02 1.08 1.00 1.17 1.20 0.183 subunit 2 P06151 L-lactate dehydrogenase A chain 0.97 0.89 1.05 1.16 1.06 1.28 1.20 0.086 Q61207 Sulfated glycoprotein 1 0.94 0.84 1.05 1.14 1.07 1.21 1.21 0.057 P16460 Argininosuccinate synthase 1.02 0.93 1.13 1.27 1.16 1.40 1.25 0.038 Q3THE2 Myosin regulatory light chain MRLC2 1.13 1.03 1.24 1.46 1.25 1.70 1.29 0.117 Q8VBT2 L-serine dehydratase 1.02 0.91 1.15 1.37 1.13 1.67 1.34 0.183 Q05816 Fatty acid-binding protein, epidermal 1.17 0.96 1.43 2.10 1.69 2.60 1.79 0.005

All values are expressed relative to a wild type control mouse (WT1). Proteins listed were significantly different in the null mice compared with wild type controls (Benjamini–Hochberg; p≤0.2).

not amenable to 2DE gel analysis, e.g. membrane bound, low 3.3. Identification of putative antioxidant response abundance or incompatible pKa values. Only two of the elements (ARE) and ARE-related motifs in the promoters identified proteins, glutathione S-transferases Mu1 and Pi1, of the Nrf2-regulated genes were shown to be Nrf2-regulated in both the 2DE analysis and the two iTRAQ analyses (Table 7). A summary of the overlap Each of the genes encoding proteins identified as Nrf2-regulated between the three different analyses is provided by Venn was interrogated for ARE or ARE-like enhancer elements in their diagram in Fig. 5. promoter regions. Using a string-based searching algorithm JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631 1623

− − Fig. 2 – Panther functional classification of proteins shown to be differentially regulated in the Nrf2( / )mouse model.

with the input term RTGABNNNTCA (representing the consen- value of 1.25 for those shown to be Nrf2 regulated. The complete sus sequence derived by Nioi et al. [44]), several ARE sequences data set for promoter analysis of all proteins significantly were identified across the panel of genes shown to be altered before correction for multiple testing is given in − − differentially expressed between Nrf2(+/+) and Nrf2( / ) mice. Table 1 of the on-line supplementary data. Table 8 shows the number of consensus sequences identified in For the matrix analysis, the patser algorithm assigns a score the 2000 bp promoter regions of the genes encoding nine foreachregionwithinthepromoterthatmatchesthe representative proteins whose expression showed the greatest position-specific probability matrix. The score is based on differential expression (>0.4-fold difference) between the two the degree of similarity to the most frequently observed mouse strains. It is apparent that there is little correlation sequence within a series of known Nrf2 target genes [45].To between the number of perfect AREs identified and the fold- define a reference score, promoters for all the 769 quantified change in expression. Indeed, the average number of consensus proteins were searched for putative ARE sequences. The mean ARE sequences identified by string-based searching across the patser score across the entire panel of genes was 2.50 whilst entire panel of proteins identified was 1.21 compared with a that for the Nrf2-regulated genes was 2.03.

Table 5 – Metacore network analysis of data from iTRAQ analysis 1. Pathway name Negative log p Number of pathway value objects

1 n-6 Polyunsaturated fatty acid biosynthesis 2.52 5 2 n-3 Polyunsaturated fatty acid biosynthesis 2.52 5 3 Regulation of lipid metabolism_Regulation of lipid metabolism via LXR, NF-Y and SREBP 2.44 3 4 Vitamin E (alfa-tocopherol) metabolism 1.98 5 5 Regulation of metabolism_Bile acids regulation of glucose and lipid metabolism via FXR 1.89 4 6 Fatty Acid Omega Oxidation 1.64 4 7 Peroxysomal straight-chain fatty acid beta-oxidation 1.64 4 8 CFTR-dependent regulation of ion channels in Airway Epithelium (norm and CF) 1.62 2 9 Cell cycle_Role of SCF complex in cell cycle regulation 1.62 2 10 Glutathione metabolism/Rodent version 1.3 5

Proteins identified in iTRAQ analysis 1 as being differentially expressed (Benjamini-Hochberg p≤0.2) were interrogated for pathway perturbation using the pathway analysis software Metacore. The total list of all quantified proteins was applied as a background for the analysis. 1624 JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631

individual hepatotoxins. Therapeutically, it is important to understand how each mechanism contributes to the cellular defence process since the Nrf2/Keap1 system provides a potential focus for the development of therapeutic strategies for management of drug or chemical-induced liver pathologies. In most of the studies conducted with hepatotoxins in Nrf2 null mice, the chemical was administered either acutely or over a short dosing period. For example, loss of protection against liver damage from paracetamol can be observed in Nrf2 null mice following a single hepatotoxic dose. Although we have previously shown that paracetamol can activate hepatic Nrf2, even at doses that do not give rise to overt toxicity [50], it seems unlikely that such transcriptional activation, and the consequent up-regulation of cellular defence proteins, could occur sufficiently rapidly to afford protection against a massive acute chemical insult. Overt liver damage can be seen within 5 h following a toxic dose of paracetamol in mice: a timescale inconsistent with the up- regulation of proteins involved in the defence response [51]. Whilst the induction of Nrf2 may play a role in damage limitation and repair, it seems likely that constitutive differences between the wild type and knockout animals represent the major factor in protecting against the initial hepatotoxic response. Although several transcriptomic studies incorpo- rate a comparison between Nrf2 wild type and null mice, none has specifically addressed the differences at the constitutive level. Consequently, this study represents the first compre- hensive global analysis of the role of Nrf2 in the basal regulation of proteins in the liver. The main aim of the study was to identify protein networks that are perturbed at the constitutive level in Nrf2 null mice compared with wild type controls. In addition, we attempted to produce a list of differentially expressed proteins that could be employed as definitive indices of Nrf2 activity and, thus, provide – Fig. 3 Western immunoblot of ATP-citrate lyase in liver a pool of potential biomarkers for application in preclinical drug (−/−) (+/+) homogenate from Nrf2 and Nrf2 mice. (A) Immunoblot for safety assessment. Such potential biomarkers might ultimately (−/−) ATP-citrate lyase in liver homogenate from four Nrf2 mice provide a translational bridge for assessment of Nrf2 activity in – (+/+) – (KO1 KO4) and four Nrf2 mice (WT1 WT4). The molecular man. By using an experimental approach incorporating three mass of ATP-citrate lyase is approximately 120 kDa. (B) Ponceau independent sample cohorts we identified twenty proteins that protein stain of the transfer membrane shown in A) indicating were Nrf2-regulated in at least two of the three independent approximately equal loading across the gel. Lane KO1 shows analyses. Of these, twelve proteins were down-regulated in Nrf2 slightly decreased loading which is consistent with the lower null mice, seven of which are involved in drug metabolism and, level of ATP-citrate lyase in the blot above. (C) Densitometric predominantly, phase II metabolism. The reproducible and analysis of immunoblot showing a statistically significant substantial reduction of proteins such as glutathione S-trans- p t (−/−) ( <0.05; Student's -test)elevationofexpressionintheNrf2 ferases mu and pi, and the UDP-glucuronosyl transferase 2B5 in − − mice compared with the wild type controls. Nrf2( / ) mice clearly indicates that protection against chemical toxins that undergo bioactivation to chemically reactive species, such as electrophiles, may be severely compromised due to lack of Nrf2 under basal conditions. The constitutive deficiency in 4. Discussion such protective proteins will almost certainly play a part in the enhanced susceptibility to chemical toxins seen upon acute Nrf2 deficient mice are highly susceptible to liver damage administration. evoked by a range of chemical hepatotoxins [25,30–34]. The Somewhat surprisingly, many proteins that were signifi- cause of this predisposition may be either reduced constitu- cantly different between the null and wild-type animals were tive expression of Nrf2-regulated genes or the loss of ability to up-regulated in the absence of Nrf2, suggesting a negative respond to the noxious stimulus by up-regulation of cellular regulation of their expression. Inspection of these proteins defence proteins. It is likely that both of these potential indicated that the majority were primarily involved in lipid mechanisms plays a role in counteracting the damage caused metabolism. Consequently, an attempt was made to categorise by exposure to chemical toxins, however, as yet the relative the differentially expressed proteins with respect to biochem- importance of each pathway has not been established for ical function by pathway analysis. MetaCore. These analyses JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631 1625

Fig. 4 – 2D gel electropherograms of Nrf2 null and wild type mouse liver proteins. (A) Representative 2DE gel of liver homogenate prepared from a wild type mouse annotated with the spot reference numbers of proteins that were found to be regulated by − − Nrf2. (B) Expanded montages of differentially expressed protein spots from Nrf2(+/+) and Nrf2( / ) mouse liver homogenates. identified lipid metabolic pathways as being highly overrepre- seven of the ten pathways shown to be significantly perturbed − − sented within the lists of significantly altered proteins when in the Nrf2( / ) mice related to the regulation of lipid biochem- compared against the entire list of proteins identified. Indeed, istry — in particular with respect to lipogenesis. 1626 JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631

Table 6 – Proteins regulated by Nrf2 identified by 2DE analysis. Protein SwissProt Protein Mr/pI Normalized Fold change p spot acc. No. spot Intensity (% total spot intensity)

− − − − Nrf2(+/+) Nrf2( / ) (Nrf2( / )/ Nrf2(+/+))

1 P24549 Aldehyde dehydrogenase family 1, subfamily A1 54468/7.91 0.19±0.01 0.15±0.01 0.78 0.0025 2 P10649 Glutathione S-transferase, mu 1 25970/7.7 0.88±0.20 0.53±0.07 0.60 0.0158 3 P30115 Glutathione S-transferase, alpha 3 25361/8.76 0.85±0.09 0.67±0.02 0.80 0.0098 4a P19157 Glutathione S-transferase, pi 1 23609/7.69 2.06±0.17 1.20±.50 0.58 0.0170 4b P19157 Glutathione S-transferase, pi 1 23609/7.69 0.29±0.03 0.017±0.05 0.57 0.0040 4c P19157 Glutathione S-transferase, pi 1 23609/7.69 0.21±0.01 0.11±0.05 0.50 0.0073 5 Q923D2 Biliverdin reductase B 22197/6.49 0.13±0.01 0.010±0.01 0.83 0.0155 6 P11588 Major urinary protein 6 20648/5.0 0.46±0.06 0.17±0.05 0.36 0.0002

Proteins from the livers of individual mice were separated by 2-DE as described in the Materials and methods. The protein spots were quantified from colloidal Coomassie blue-stained gels using ImageMasterTM 2D Elite software. Spot intensities were normalized to the total spot intensity for each gel and expressed as the mean percentage value±SD (n=4 for each group). Proteins that were significantly different (Student's t-test; p<0.05) between the wild type and Nrf2 null mice are shown.

A potential role for Nrf2 in the regulation of lipid biochem- profiles. Both studies noted that a high proportion of the up- istry, and more specifically in the disposition of fatty acids, has regulated mRNAs coded for proteins involved in lipid homeo- only recently been recognised [48,49,52]. This probably reflects stasis. In the former study, 4 weeks on a high fat diet resulted the fact that earlier studies with transgenic animals predom- in a marked increase in the mRNAs for several cholesterol inantly concentrated on proteins that are directly correlated synthetic and up-take genes, including LDL receptor, HMGcoA with Nrf2 activity, which comprises principally proteins reductase, HMGCoA synthase and SR-B1. Interestingly, the involved in cellular antioxidant defence. Three recent studies mRNA for Nrf2 itself was substantially reduced following this have, however, noted the reciprocal relationship between Nrf2 diet, suggesting that the Keap1/Nrf2 pathway may be directly function and the expression of multiple lipid-related gene regulated by certain dietary lipids. This must be balanced − − products. Studies by Tanaka et al. [53],involvingNrf2( / ) mice against the fact that some terpenoids, which are also lipids and fed a high fat diet, and Yates et al. [54], which compared Keap1 share synthetic pathways with cholesterol, are among the knockout mice with mice exposed to a potent Nrf2 inducer, most potent activators of Nrf2 in mouse models [55,56].For utilized transcriptomic approaches to define gene expression example, the synthetic triterpenoid CDDO-Im has been shown

Table 7 – Proteins identified as Nrf2 dependent in two or more analyses. SwissProt Protein name iTRAQ Analysis 1 iTRAQ Analysis 2 2DE gel analysis acc. no. Fold-change p Fold-change p Fold-change p

Q8VCW8 Acyl-CoA synthetase family member 2, mitochondrial 0.83 0.030 0.86 0.132 P31786 Acyl-CoA-binding protein 1.26 0.073 1.18 0.207 Q91VA0 Acyl-coenzyme A synthetase ACSM1, mitochondrial 0.78 0.039 0.81 0.001 Q9JII6 Alcohol dehydrogenase [NADP+] 0.90 0.121 0.85 0.003 P24549 Aldehyde dehydrogenase family 1, subfamily A1 0.76 0.127 0.72 0.021 0.78 0.003 P16460 Argininosuccinate synthase 1.26 0.162 1.25 0.038 Q91X77 Cytochrome P450 2C50 0.65 0.162 0.58 0.001 Q05816 Fatty acid-binding protein, epidermal 2.97 0.009 1.79 0.005 Q9QXD6 Fructose-1,6-bisphosphatase 1 1.14 0.144 1.18 0.117 P30115 Glutathione S-transferase, alpha 3 0.70 0.001 0.80 0.010 P10649 Glutathione S-transferase, mu 1 0.42 0.009 0.44 0.001 0.60 0.016 P19157 Glutathione S-transferase, pi 1 0.55 0.011 0.62 0.001 0.55 0.009 Q8VCC2 Liver carboxylesterase 1 0.58 0.042 0.56 0.001 P06151 L-lactate dehydrogenase A chain 1.18 0.178 1.20 0.086 Q8VBT2 L-serine dehydratase 1.51 0.078 1.34 0.183 P02762 Major urinary protein 6 0.35 0.057 0.36 0.001 Q99P30 Peroxisomal coenzyme A diphosphatase NUDT7 1.43 0.044 1.13 0.039 Q61207 Sulfated glycoprotein 1 1.22 0.121 1.21 0.057 O70475 UDP-glucose 6-dehydrogenase 0.81 0.061 0.73 0.183 P17717 UDP-glucuronosyltransferase 2B5 0.55 0.004 0.55 0.001

− − Each protein was significantly (p<0.05, Student t-test) over- or underexpressed in Nrf2( / ) mice compared with the wild type controls in at least two out of the three independent. Fold changes are the ratios of the mean expression changes from 4 to 6 mice. JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631 1627

was seen with epidermal fatty acid binding protein (FABP5; 2.97). MUP6 is a member of a species and sex specific class of secreted proteins synthesised in the liver but used by male mice for a variety of behavioural purposes, including territorial marking and mate attraction [61]. Major urinary proteins (MUPs) are lipid binding molecules that are specifically tailored to the transport of pheromones: following urinary excretion the pheromones are released only slowly from the MUP to prolong their signalling properties. Curiously, fatty acid binding proteins belong to the same class of proteins as the MUPs — both are lipocalins — and they fulfil a similar function, both being involved in lipid transport. Consequently, the two proteins whose expression was most disparately Fig. 5 – Venn diagram indicating the overlap between the affected by loss of Nrf2 were of a similar type, again proteins identified as Nrf2-regulated across the three different emphasising the complexity of Nrf2 regulation of lipid analyses. metabolism. It is possible that up regulation of FABP5 occurred at the expense of MUP6 synthesis, which then showed a marked fall in expression; further work is required to to reduce hepatic lipid accumulation in mice on a high fat diet understand the interaction between these proteins and the through activation of Keap1/Nrf2 signalling [49]. Clearly, the precise role of Nrf2. Perturbation of the expression of FABPs role of Nrf2 in lipid homeostasis is complex and requires and MUPs has been shown in other mouse models involving further clarification. One of the most recent demonstrations antioxidant proteins. CuZn superoxide dismutase (Sod1) that Nrf2 is negatively linked to the expression of lipid-related deficient mice showed a marked decrease in MUP11 and genes results from a study by Chowdary et al. [48] investigating MUP8 expression, but in this case FABP1 was also down- the influence of Nrf2 on non-alcoholic hepatosteatosis (NASH). regulated [62]. This disease is characterised by macro- and/or micro-vesicular A requisite property of all genes identified thus far to be vacuolization of hepatocytes and can be induced by adminis- Nrf2-responsive is that they contain an ARE sequence. tering a methionine/choline deficient diet. The histopatholog- Consequently, it was important to seek the presence of such ical symptoms of the condition were exacerbated in Nrf2 AREs within the promoter regions of the genes encoding the deficient mice and was accompanied by up-regulation of identified proteins, particularly those not previously reported proteins involved in lipid metabolism including Adrp, a fatty to be Nrf2-dependent. This was accomplished using software acid- and cholesterol-binding protein that promotes accumu- available in the public domain that allows both multiple string lation of triacylglycerols and stimulates the uptake of fatty searching and pattern recognition analysis of 5′-flanking acids [57]. regions. String searching based on the ‘perfect’ ARE sequence One of the proteins involved in lipid metabolism shown to of RTGABNNNGA as defined by Rushmore [63–66] indicated − − be strongly enhanced in Nrf2( / ) mice in the current study was that a wide variety of potential AREs was present across the 16 ATP citrate synthase (also known as ATP-citrate lyase). The genes significantly altered in Nrf2 null animals. However, only almost two-fold increase in this enzyme indicated by the GSTM1 among the most highly Nrf2 regulated genes had more initial iTRAQ analysis (Table 2) was confirmed by Western than two perfect ARE sequences within their promoter immunoblotting (Fig. 3). As far as we are aware, regulation of regions. Overall, there was little difference in the average this protein by Nrf2 has not been previously demonstrated. number of AREs found in the Nrf2-regulated genes ATP-citrate lyase plays a critical role in acetyl CoA production (mean=1.25) compared with the number found right across within the cytoplasm of most cells, and is especially active in the pool of identified proteins (1.21). Analysis based on a liver [58]. In the presence of ATP and Coenzyme A, ATP-citrate pattern recognition algorithm (patser) similarly showed little lyase is able to cleave citrate to form acetyl-CoA and difference between up-or down-regulated genes when com- oxaloacetic acid. The latter is a precursor for pyruvate which pared against the total protein pool. These results suggest that sits at the crossroads of multiple biochemical pathways, such prediction of Nrf2-regulated genes based on regulatory as amino acid synthesis, glycogenolysis and lipogenesis. sequence analysis may be an unreliable approach and a Furthermore, it has recently been shown that ATP-citrate molecular-based promoter analysis is required to define the lyase is a key enzyme in the acetylation of histones, and may precise site of Nrf2 activation. therefore play a major role in gene transcription [59,60]. The In summary, this study has identified a panel of Nrf2- demonstration that loss of Nrf2 results in such a large up- dependent hepatic proteins that is statistically robust and regulation of this already abundant protein may therefore demonstrated both decreased and enhanced expression in the have significant implication for multiple cellular functions, absence of the Nrf2 gene. Twenty proteins were identified as and this is the subject of further investigation in our Nrf2-regulated in at least two of the three analyses providing laboratories. further confidence that these proteins might provide useful Of all the statistically significant changes in protein candidate biomarkers for future translational studies. The expression between wild type and Nrf2 null mice, numerically number of proteins that can be interrogated by currently the largest fold decrease was seen with major urinary protein available proteomic technology is far fewer than the number 6 (MUP6; 0.35 relative to wild type) whilst the biggest increase of genes interrogated in the oligoarray studies carried out by 1628

Table 8 – Promoter analysis for the mouse genes encoding Nrf2-regulated proteins. String search Matrix analysis (patser) Highest scoring ARE (dna-pattern)

SwissProt Protein Fold- Number of Number Highest Mean SD Location Sequence 1612 (2010) 73 PROTEOMICS OF JOURNAL acc. no. name change consensus of score score from to sequences matching (RTGABNNNGCA) sequences

P02762 Major urinary protein 6 0.35 0 14 4.89 2.03 1.07 −1935 −1923 ttccCTGTCACTAAGCAtgtt P10649 Glutathione S-transferase Mu 1 0.41 4 15 4.40 2.42 1.09 −56 −44 gtggGCAGGACAAAACAgcgg P19157 Glutathione S-transferase P 1 0.54 0 13 4.02 2.11 0.98 −68 −56 aacgTGTTGAGTCAGCAtccg Q91WG8 Bifunctional UDP-N-acetylglucosamine 2-epimerase/N- 0.55 0 12 5.95 2.50 1.70 −387 −375 gcagGGGTGGCAAAGCTtaaa acetylmannosamine kinase P17717 UDP-glucuronosyltransferase 2B5 0.55 1 13 5.59 2.40 1.23 −398 −386 cagtCCATGACTGAGTTtgaa Q99P30 Peroxisomal coenzyme A diphosphatase NUDT7 1.41 1 8 4.68 2.49 1.14 −848 −836 caagGCATTACACAGCCcagg Q8JZR0 Long-chain-fatty-acid-CoA ligase 5 1.57 1 10 7.66 2.56 1.90 −1234 −1222 cttaGAATGACCCAGCCcttg Q91V92 ATP-citrate synthase 1.75 1 9 10.02 3.26 2.58 −1899 −1887 agaaAAATGACTAAGCAggta Q8VCH0 3-ketoacyl-CoA thiolase B, peroxisomal 2.21 2 15 5.84 2.55 1.44 −137 −125 tgggGGAAGACTCAGGAagag Q05816 Fatty acid-binding protein, epidermal 2.81 0 15 4.37 2.59 0.86 −1728 −1716 agtgGGATGTCGCAGCTcagg Mean values for all Nrf2-regulated proteins 1.26 1.25 13.69 5.62 2.50 1.33 Mean values for all down-regulated Nrf2-dependent proteins 0.57 1.00 15.40 5.20 2.54 1.23 –

Mean values for all up-regulated Nrf2-dependent proteins 1.57 1.36 12.91 5.81 2.49 1.37 1631 Mean values for all proteins identified 1.21 13.20 6.48 2.03 1.62

Sequences of the genes of Nrf2-regulated proteins were obtained from the ENSMBL mouse genome database and interrogated for ARE and ARE/like consensus sequences using the RSAT analysis software (http://rsat.ulb.ac.be/rsat/). Both string-based (dna-pattern) and matrix-based (patser) pattern searching strategies were adopted (see text for details). For the dna-pattern analysis, returned sequences were rated against the ‘perfect’ consensus sequence RTGABNNNGCA. For the patser analysis, the number of sequences matching the position specific scoring matrix with a score >1 are given, along with the highest score attained. For comparison, equivalent data from the entire set of identified proteins is included at the foot of the table. JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631 1629 various groups [37–39]. This reflects, at least in part, the much REFERENCES lower analyte coverage of proteomic approaches and the fact that enhanced mRNA levels are often not mirrored by an [1] Ames BN. Dietary carcinogens and anticarcinogens. Oxygen equivalent up-regulation of the corresponding protein [41]. radicals and degenerative diseases. Science 1983;221:1256–64. From a global perspective, there was little concordance between [2] Miller JA. Carcinogenesis by chemicals: an overview–G. H. A. the proteins identified here and in all the various microarray Clowes memorial lecture. Cancer Res 1970;30:559–76. studies presented previously. However, there was some overlap [3] Sims P, Grover PL, Swaisland A, Pal K, Hewer A. Metabolic with well-established Nrf2-regulated mouse genes, such as activation of benzo(a)pyrene proceeds by a diol-epoxide. – glutathione transferases and glucuronyl transferases. Overall, Nature 1974;252:326 8. [4] Prestera T, Zhang Y, Spencer SR, Wilczak CA, Talalay P. The the number of phase II proteins shown in the study to be Nrf2- electrophile counterattack response: protection against regulated was surprisingly small and only one pathway neoplasia and toxicity. Adv Enzyme Regul 1993;33:281–96. (glutathione metabolism) involved in cellular defence was [5] Primiano T, Sutter TR, Kensler TW. Antioxidant-inducible identified as Nrf2 regulated from the MetaCore analysis. genes. Adv Pharmacol 1997;38:293–328. iTRAQ suffers from the same limitation as all global proteomic [6] Buetler TM, Gallagher EP, Wang C, Stahl DL, Hayes JD, Eaton methodologies in that, although it can sample around 1000 DL. Induction of phase I and phase II drug-metabolizing enzyme mRNA, protein, and activity by BHA, ethoxyquin, and proteins simultaneously, high abundance species still dom- oltipraz. Toxicol Appl Pharmacol 1995;135:45–57. inate. Furthermore, the inherent variability associated with [7] Hayes JD, Pulford DJ. The glutathione S-transferase supergene any technique that relies on the quantification of a protein family: regulation of GST and the contribution of the based on the analysis of multiple peptide mass spectra isoenzymes to cancer chemoprotection and drug resistance. (where the number of peptides identified may vary between Crit Rev Biochem Mol Biol 1995;30:445–600. analyses) means that fold changes less than 20% are unlikely [8] Alam J, Stewart D, Touchard C, Boinapally S, Choi AM, Cook JL. to yield statistically robust quantitative data. In biological Nrf2, a Cap'n'Collar transcription factor, regulates induction of the heme oxygenase-1 gene. J Biol Chem 1999;274:26071–8. terms, however, a 10–20% decrease in protein expression [9] Chanas SA, Jiang Q, McMahon M, McWalter GK, McLellan LI, could have a marked bearing on the toxicological conse- Elcombe CR, et al. Loss of the Nrf2 transcription factor causes quence of chemical exposure. However, the strength of such a marked reduction in constitutive and inducible expression global approaches is that they can reveal hitherto unrecog- of the glutathione S-transferase Gsta1, Gsta2, Gstm1, Gstm2, nised roles for key cellular regulators, as evidenced here by Gstm3 and Gstm4 genes in the livers of male and female mice. the demonstration that lipid homeostasis is a key function of Biochem J 2002;365:405–16. Nrf2. In conclusion, this study has identified a reliable panel [10] Hayes JD, Chanas SA, Henderson CJ, McMahon M, Sun C, Moffat GJ, et al. The Nrf2 transcription factor contributes both of proteins that are reproducibly Nrf2-regulated at the to the basal expression of glutathione S-transferases in constitutive level. This paves the way for the future transla- mouse liver and to their induction by the chemopreventive tional applications in the pursuit of a human biomarker that synthetic antioxidants, butylated hydroxyanisole and would enable the assessment of the role of Nrf2 in man. The ethoxyquin. Biochem Soc Trans 2000;28:33–41. identification of two principal sets of seemingly unrelated [11] Ishii T, Itoh K, Takahashi S, Sato H, Yanagawa T, Katoh Y, proteins — one involved in cellular defence, the other in lipid et al. Transcription factor Nrf2 coordinately regulates a group of oxidative stress-inducible genes in macrophages. J Biol homeostasis — that are reciprocally regulated by Nrf2 was an Chem 2000;275:16023–9. unexpected finding of this study, and may provide a powerful [12] Kim YC, Masutani H, Yamaguchi Y, Itoh K, Yamamoto M, reservoir of diagnostic biomarkers for development in both Yodoi J. Hemin-induced activation of the thioredoxin gene by animal and human studies of Nrf2 activity. Nrf2. A differential regulation of the antioxidant responsive element by a switch of its binding factors. J Biol Chem 2001;276:18399–406. Acknowledgements [13] Lee JM, Moehlenkamp JD, Hanson JM, Johnson JA. Nrf2-dependent activation of the antioxidant responsive element by tert-butylhydroquinone is independent of This work was supported by The Wellcome Trust. Azman oxidative stress in IMR-32 human neuroblastoma cells. Abdullah would like to thank the Ministry of Science, Technol- Biochem Biophys Res Comm 2001;280:286–92. ogy and Innovation of the Government of Malaysia as well as the [14] Nguyen T, Huang HC, Pickett CB. Transcriptional regulation of National University of Malaysia for sponsoring his doctoral the antioxidant response element. Activation by Nrf2 and – studies at the University of Liverpool. JW and RS are BBSRC- repression by MafK. J Biol Chem 2000;275:15466 73. funded PhD students with further support from AstraZeneca [15] Wild AC, Moinova HR, Mulcahy RT. Regulation of gamma-glutamylcysteine synthetase subunit gene Pharmaceuticals. We are grateful to Brian Lane (Department of expression by the transcription factor Nrf2. J Biol Chem Medical Microbiology, School of Infection and Host Defence, 1999;274:33627–36. University of Liverpool) and to Tracey Farragher (Centre for [16] Favreau LV, Pickett CB. The rat quinone reductase antioxidant Medical Statistics and Health Evaluation, University of Liver- response element. Identification of the nucleotide sequence pool) for the statistical advice and support. required for basal and inducible activity and detection of antioxidant response element-binding proteins in hepatoma and non-hepatoma cell lines. J Biol Chem 1995;270:24468–74. [17] Rushmore TH, King RG, Paulson KE, Pickett CB. Regulation of Appendix A. Supplementary data glutathione S-transferase Ya subunit gene expression: identification of a unique xenobiotic-responsive element Supplementary data associated with this article can be found, controlling inducible expression by planar aromatic in the online version, at doi:10.1016/j.jprot.2010.03.018. compounds. Proc Natl Acad Sci U S A 1990;87:3826–30. 1630 JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631

[18] Rushmore TH, Pickett CB. Transcriptional regulation of the rat [35] Kwak MK, Wakabayashi N, Itoh K, Motohashi H, Yamamoto glutathione S-transferase Ya subunit gene. Characterization M, Kensler TW. Modulation of gene expression by cancer of a xenobiotic-responsive element controlling inducible chemopreventive dithiolethiones through the Keap1-Nrf2 expression by phenolic antioxidants. J Biol Chem 1990;265: pathway. Identification of novel gene clusters for cell 14648–53. survival. J Biol Chem 2003;278:8135–45. [19] Mulcahy RT, Wartman MA, Bailey HH, Gipp JJ. Constitutive [36] Lee JM, Calkins MJ, Chan K, Kan YW, Johnson JA. and beta-naphthoflavone-induced expression of the human Identification of the NF-E2-related factor-2-dependent genes gamma-glutamylcysteine synthetase heavy subunit gene is conferring protection against oxidative stress in primary regulated by a distal antioxidant response element/TRE cortical astrocytes using oligonucleotide microarray analysis. sequence. J Biol Chem 1997;272:7445–54. J Biol Chem 2003;278:12029–38. [20] Alam J, Wicks C, Stewart D, Gong P, Touchard C, Otterbein S, [37] Abbott A. A post-genomic challenge: learning to read patterns et al. Mechanism of heme oxygenase-1 gene activation by of protein synthesis. Nature 1999;402:715–20. cadmium in MCF-7 mammary epithelial cells. Role of p38 [38] Bradford MM. A rapid and sensitive method for the kinase and Nrf2 transcription factor. J Biol Chem 2000;275: quantitation of microgram quantities of protein utilizing the 27694–702. principle of protein-dye binding. Anal Biochem 1976;72: [21] Orino K, Lehman L, Tsuji Y, Ayaki H, Torti SV, Torti FM. 248–54. Ferritin and the response to oxidative stress. Biochem J [39] Gorg A, Obermaier C, Boguth G, Harder A, Scheibe B, 2001;357:241–7. Wildgruber R, et al. The current state of two-dimensional [22] Yoh K, Itoh K, Enomoto A, Hirayama A, Yamaguchi N, electrophoresis with immobilized pH gradients. Kobayashi M, et al. Nrf2-deficient female mice develop Electrophoresis 2000;21:1037–53. lupus-like autoimmune nephritis. Kidney Int 2001;60:1343–53. [40] Courchesne PL, Patterson SD. Identification of proteins by [23] Chan K, Lu R, Chang JC, Kan YW. NRF2, a member of the NFE2 matrix-assisted laser desorption/ionization mass family of transcription factors, is not essential for murine spectrometry using peptide and fragment ion masses. erythropoiesis, growth, and development. Proc Natl Acad Sci Methods Mol Biol 1999;112:487–511. U S A 1996;93:13943–8. [41] Fountoulakis M, Berndt P, Boelsterli UA, Crameri F, Winter M, [24] Chan K, Han XD, Kan YW. An important function of Nrf2 in Albertini S, et al. Two-dimensional database of mouse liver combating oxidative stress: detoxification of acetaminophen. proteins: changes in hepatic protein levels following Proc Natl Acad Sci U S A 2001;98:4611–6. treatment with acetaminophen or its nontoxic regioisomer [25] Enomoto A, Itoh K, Nagayoshi E, Haruta J, Kimura T, O'Connor 3-acetamidophenol. Electrophoresis 2000;21:2148–61. T, et al. High sensitivity of Nrf2 knockout mice to [42] Fountoulakis M, Juranville JF, Berndt P, Langen H, Suter L. acetaminophen hepatotoxicity associated with decreased Two-dimensional database of mouse liver proteins. An expression of ARE-regulated drug metabolizing enzymes and update Electrophoresis 2001;22:1747–63. antioxidant genes. Toxicol Sci 2001;59:169–77. [43] van Helden J. Regulatory sequence analysis tools. Nucleic [26] Itoh K, Chiba T, Takahashi S, Ishii T, Igarashi K, Katoh Y, et al. Acids Res 2003;31:3593–6. An Nrf2/small Maf heterodimer mediates the induction of [44] Nioi P, McMahon M, Itoh K, Yamamoto M, Hayes JD. phase II detoxifying enzyme genes through antioxidant Identification of a novel Nrf2-regulated antioxidant response response elements. Biochem Biophys Res Commun 1997;236: element (ARE) in the mouse NAD(P)H: quinone 313–22. oxidoreductase 1 gene: reassessment of the ARE consensus [27] Martin F, van Deursen JM, Shivdasani RA, Jackson CW, sequence. Biochem J 2003;374:337–48. Troutman AG, Ney PA. Erythroid maturation and globin gene [45] Nerland DE. The antioxidant/electrophile response element expression in mice with combined deficiency of NF-E2 and motif. Drug Metab Rev 2007;39:235–48. nrf-2. Blood 1998;91:3459–66. [46] R_Development_Core_Team. R: A language and environment [28] McMahon M, Itoh K, Yamamoto M, Chanas SA, Henderson CJ, for statistical computing. R Foundation for Statistical McLellan LI, et al. The cap 'n' collar basic leucine zipper Computing, Vienna, Austria; 2009. transcription factor Nrf2 (NF-E2 p45-related factor 2) controls [47] Katz MH. Multivariable analysis: a primer for readers of both constitutive and inducible expression of intestinal medical research. Ann Intern Med 2003;138:644–50. detoxification and glutathione biosynthetic enzymes. Cancer [48] Chowdhry S, Nazmy MH, Meakin PJ, Dinkova-Kostova AT, Res 2001;61:3299–307. Walsh SV, Tsujita T, et al. Loss of Nrf2 markedly exacerbates [29] Thimmulappa RK, Mai KH, Srisuma S, Kensler TW, Yamamato non-alcoholic steatohepatitis. Free Radic Biol Med 2009. M, Biswal S. Identification of Nrf2-regulated genes induced by [49] Shin S, Wakabayashi J, Yates MS, Wakabayashi N, Dolan PM, the chemopreventive agent sulforaphane by oligonucleotide Aja S, et al. Role of Nrf2 in prevention of high-fat diet-induced microarray. Cancer Research 2002;62:5196–203. obesity by synthetic triterpenoid CDDO-Imidazolide. Eur J [30] Chan KM, Han XD, Kan YW. An important function of Nrf2 in Pharmacol 2009;620:138–44. combating oxidative stress: Detoxification of acetaminophen. [50] Goldring CE, Kitteringham NR, Elsby R, Randle LE, Clement Proc Natl Acad Sci U S A 2001;98:4611–6. YN, Williams DP, et al. Activation of hepatic Nrf2 in vivo by [31] Xu WH, Hellerbrand C, Kohler UA, Bugnon P, Kan YW, Werner S, acetaminophen in CD-1 mice. Hepatology 2004;39:1267–76. et al. The Nrf2 transcription factor protects from toxin-induced [51] Kitteringham NR, Powell H, Clement YN, Dodd CC, Tettey JNA, liver injury and fibrosis. Lab Invest 2008;88:1068–78. Pirmohamed M, et al. Hepatocellular response to chemical [32] Lu Y, Gong P, Cederbaum AI. Pyrazole induced oxidative liver stress in CD-1 mice: Induction of early genes and injury independent of CYP2E1/2A5 induction due to Nrf2 gamma-glutamylcysteine synthetase. Hepatology 2000;32: deficiency. Toxicology 2008;252:9–16. 321–33. [33] Lamle J, Marhenke S, Borlak J, Von Wasielewski R, Eriksson [52] Sugimoto H, Okada K, Shoda J, Warabi E, Ishige K, Ueda T, et al. CJP, Geffers R, et al. Nuclear factor-eythroid 2-related factor 2 Deletion of nuclear factor-E2-related factor-2 leads to rapid prevents alcohol-induced fulminant liver injury. onset and progression of nutritional steatohepatitis in mice. Gastroenterology 2008;134:1159–68. Am J Physiol Gastrointest Liver Physiol 2010;298:G283–94. [34] Umemura T, Kuroiwa Y, Kitamura Y, Ishii Y, Kanki K, Kodama [53] Tanaka Y, Aleksunes LM, Yeager RL, Gyamfi MA, Esterly N, Y, et al. A crucial role of Nrf2 in in vivo defense against Guo GL, et al. NF-E2-related factor 2 inhibits lipid oxidative damage by an environmental pollutant, accumulation and oxidative stress in mice fed a high-fat diet. pentachlorophenol. Toxicol Sci 2006;90:111–9. J Pharmacol Exp Ther 2008;325:655–64. JOURNAL OF PROTEOMICS 73 (2010) 1612– 1631 1631

[54] Yates MS, Tran QT, Dolan PM, Osburn WO, Shin S, McCulloch [60] Cousins RJ, Aydemir TB, Lichten LA. Plenary Lecture 2 CC, et al. Genetic versus chemoprotective activation of Nrf2 Transcription factors, regulatory elements and nutrient-gene signaling: overlapping yet distinct gene expression profiles communication. Proc Nutr Soc 2009:1–4. between Keap1 knockout and triterpenoid-treated mice. [61] Beynon RJ, Hurst JL. Urinary proteins and the modulation of Carcinogenesis 2009;30:1024–31. chemical scents in mice and rats. Peptides 2004;25:1553–63. [55] Liu XP, Goldring CEP, Wang HY, Copple IM, Kitteringham NR, [62] Elchuri S, Naeemuddin M, Sharpe O, Robinson WH, Huang TT. Park BK. Extract of Ginkgo biloba induces Identification of biomarkers associated with the development glutathione-S-transferase subunit-P1 in vitro. Phytomedicine of hepatocellular carcinoma in CuZn superoxide dismutase 2009;16:451–5. deficient mice. Proteomics 2007;7:2121–9. [56] Thimmulappa RK, Fuchs RJ, Malhotra D, Scollick C, Traore K, [63] Ellis EM, Slattery CM, Hayes JD. Characterization of the rat Bream JH, et al. Preclinical evaluation of targeting the Nrf2 aflatoxin B-1 aldehyde reductase gene, AKR7A1. Structure pathway by triterpenoids (CDDO-Im and CDDO-Me) for and chromosomal localization of AKR7A1 as well as protection from LPS-induced inflammatory response and identification of antioxidant response elements in the gene reactive oxygen species in human peripheral blood promoter. Carcinogenesis 2003;24:727–37. mononuclear cells and neutrophils. Antioxid Redox Signal [64] Motohashi H, O'Connor T, Katsuoka F, Engel JD, Yamamoto M. 2007;9:1963–70. Integration and diversity of the regulatory network composed [57] Martin S, Parton RG. Lipid droplets: a unified view of a of Maf and CNC families of transcription factors. Gene dynamic organelle. Nat Rev Mol Cell Biol 2006;7:373–8. 2002;294:1–12. [58] Elshourbagy NA, Near JC, Kmetz PJ, Wells TN, Groot PH, Saxty [65] Rushmore TH, Morton MR, Pickett CB. The antioxidant BA, et al. Cloning and expression of a human ATP-citrate responsive element. Activation by oxidative stress and lyase cDNA. Eur J Biochem 1992;204:491–9. identification of the DNA consensus sequence required for [59] Wellen KE, Hatzivassiliou G, Sachdeva UM, Bui TV, Cross JR, functional activity. J Biol Chem 1991;266:11632–9. Thompson CB. ATP-citrate lyase links cellular metabolism to [66] Wasserman WW, Fahl WE. Functional antioxidant responsive histone acetylation. Science 2009;324:1076–80. elements. Proc Natl Acad Sci U S A 1997;94:5361–6.